In short
How Samsara applies edge AI and cloud AI to “physical world” operations—covering safety, efficiency, and compliance—using massive real-world data (20 trillion data points; 90 billion miles annually), with examples like drowsiness detection, fuel theft detection, and commercial navigation that avoids low bridges.
Guest background
Praveen Murugesan is VP of Engineering at Samsara. He previously worked at Uber (7+ years). He’s responsible for teams applying AI to interpret video, vehicle, and sensor data, and he’s based in Amsterdam during the recording.
Key claims
Safety use cases run primarily on-device (edge) for reliability/latency; non-latency problems like fuel analytics run in the cloud. Samsara’s “AI gateway” centralizes access to multiple LLMs (OpenAI, Anthropic, Gemini) while handling cost, security, compliance, and feedback loops. Road-planning automation should be assistive, not fully replacing human route planners.
Notable examples
Edge safety cameras detect driver drowsiness and run vision models on constrained hardware. Fuel theft is a ~$130B/year global problem; Samsara uses statistical models to handle noisy, variant sensor readings (e.g., dual-tank trucks). Commercial navigation is context-aware for vehicle type and enforces org policies (e.g., turn restrictions) to prevent incidents like trucks hitting low bridges. Quantum computing could improve routing/scheduling (vehicle routing/traveling salesman) by enabling more optimal solutions with real-time traffic inputs.
Written by AI. May contain mistakes. Listen to the episode to check what was said.
Chapters
Tap a time to open that second in VOUnderstanding Samsara's Data Challenge
0:41 to 1:28
Discussion about the vast data collected by Samsara and its implications.
“Praveen, welcome to the Super Data Science Podcast.”
Insights from Samsara's Engineering
1:29 to 2:24
Praveen shares insights into how AI is applied to large datasets.
“So tell us about that huge volume of data and what the challenges are, what the exciting things are about dealing with things on that scale.”
Applications of AI in Physical Operations
2:25 to 3:53
Exploration of AI applications in safety and efficiency in physical operations.
“and Samsara particularly is very interesting for the same reason that you mentioned, which is like the volume of data is just so much.”
Safety Solutions with AI Cameras
3:54 to 4:25
Discussion on AI cameras for monitoring driving behaviors and safety.
“Like how do I operate really efficiently?”
Edge Computing and AI Processing
4:26 to 6:19
Delving into the technical aspects of edge computing and AI processing in vehicles.
“Now, maybe I'll give you a couple of examples and we can even talk a little bit more of maybe deep dive a couple of them.”
Challenges in Real-Time Machine Vision
6:20 to 8:13
Exploring the challenges of implementing real-time machine vision in vehicles.
“yeah so we actually um do a like at a high level we basically do a lot of our training in our backend for these models.”
Building for Connectivity Challenges
8:14 to 10:46
Discussion on building solutions for environments with poor connectivity.
“You have to then operate with the hardware limitations, especially when it comes down to the inference models.”
Fuel Efficiency and Theft Solutions
10:47 to 14:00
Introduction to AI solutions addressing fuel efficiency and theft issues.
“Well, so yeah, so we've gone into a bit of detail now.”
Understanding Fuel Anomalies
14:00 to 14:55
Discussion on detecting fuel anomalies using statistical models.
“Like actually we do more like a statistical model here where, which can actually smooth out the variance and like effectively like improve the hit rate when we find like, you know, the anomalies in fuel.”
Global Application of AI in Fuel
15:25 to 18:04
Exploration of AI's application across various global sectors.
“So you have all of these factors feeding into the model and you were able to get kind of labeled training data in one region.”
Show all 23 chapters
Trade-offs in Edge Computing
18:04 to 21:05
Discussion on edge computing vs. cloud processing for commercial drivers.
“And so when you're doing this kind of edge computing and you're helping, say, commercial drivers operate more effectively, you know, digging a bit more maybe into that kind of cab camera example.”
AI Gateway for Engineers
21:05 to 23:00
Explanation of the AI gateway concept and its benefits for engineers.
“Right, yeah, let's talk about that a bit more.”
Defining the Physical AI Roadmap
23:00 to 28:00
Insights into how Samsara defines its AI roadmap based on customer needs.
“So the gateway kind of allows you to, it allows your engineers, your scientists, your developers to be able to take into account capability versus cost considerations right out of the box.”
Customer-Centric Approach to AI
28:00 to 29:18
Learn how a customer feedback loop drives technology development.
“Like it's usually a very good filter, I would say, like, you know, whether you're chasing randomized ideas or like something that's actually concretely solving pain.”
Exploring Quantum Computing
29:18 to 30:26
Understand the potential of quantum computing in logistics and routing.
“So yeah, so we've talked about now LLMs and the cool capabilities that all of us have access to now.”
Quantum Solutions for Complex Problems
30:26 to 34:11
Discover how quantum computing can optimize complex routing challenges.
“I think for, um, when you really think about it, right?”
Commercial Navigation Solutions
34:11 to 36:28
Explore how Samsara's navigation tools address real trucking issues.
“It's going to be interesting to see all the application areas that quantum ends up making a big difference in.”
Innovations in Route Planning
36:28 to 40:16
Learn about automated solutions transforming traditional route planning.
“I think a lot of, for them, the driving need is also compliance and safety of their drivers, right?”
Human Context in Algorithmic Planning
40:16 to 42:00
Understand the importance of integrating human context in algorithmic solutions.
“And so, you know, things like things that previously might've been done in a spreadsheet are now done algorithmically and automatically.”
Balancing Automation and Human Connection
42:00 to 44:40
Explore the balance between technological optimization and maintaining human relationships in logistics.
“So they wanted to preserve some of the relationships that they had between driver to the actual customer of theirs.”
Hiring for Success at Samsara
44:40 to 46:49
Learn about the qualities and skills that Samsara looks for in candidates as they expand.
“I'm glad that you were able to dig into that.”
Book Recommendation: A Classic Sci-Fi
46:49 to 50:11
Discover the themes of humanity and technology in Philip K. Dick's 'Do Androids Dream of Electric Sheep?'
“which means like, you know, a lot of adaptability and agility is a part of it, right?”
Wrap-Up and Reflections on AI
50:11 to 51:03
Reflect on the insights shared about AI applications and the future of technology.
“Thank you for that great recommendation, Praveen.”
Transcript
Automatic transcript. May contain errors.0:00Jon Krohn:What happens when a truck driver falls asleep at the wheel and there's no cell signal for miles? AI on the edge, designed for physical systems, is the answer. Welcome to the Super Data Science Podcast. I'm your host, Jon Krohn. Today's excellent episode features Praveen Murugesan, VP of Engineering at Samsara, a physical operations platform that processes over 20 trillion data points for the world's leading organizations across construction, transportation, manufacturing, retail, logistics, and the public sector. hear all about how some SARA engineers make AI impactful across such a broad range of physical applications in this practical and excellent episode.
0:39Jon Krohn:Enjoy. This episode of Super Data Science is made possible by Dell, Intel, Excel Data, and the Open Data Science Conference. Praveen, welcome to the Super Data Science Podcast. Great to have you on the show. Where are you calling in from today? Hey, John. Thanks for having me. I'm calling in from Amsterdam in the Netherlands. Ah, yes. Appropriately in Amsterdam, a big center of shipping for today's episode, because we're talking all about physical systems. You're one of the world's leaders in providing software for physical systems. And so we're going to talk a lot about your company, Samsara.
1:18Jon Krohn:But right before we get into that, I want to talk about the data that you have to work with, because I was blown away as we were doing research for your episode on the volumes of data that you have. So as VP of Engineering at Samsara, you're responsible for overseeing teams that apply AI to interpret massive volumes of video, vehicle and sensor data, 20 trillion data points covering 90 billion miles annually. So tell us about that huge volume of data and what the challenges are, what the exciting things are about dealing with things on that scale. No, it's a great question. I think it's funny that actually Sanjadar, founder, recently mentioned that, hey, we probably are collecting more real world data than anybody else.
2:04And I think maybe Uber was the only one that's probably collecting a little bit more.
2:09Jon Krohn:And you actually, you previously worked at Uber. I did, actually. It was a pretty good journey, like, you know, seven plus years. I think for me, my time at Uber actually like made me like want to keep working on more physical world problems. It's one of the reasons like I kind of stuck to the space. and Samsara particularly is very interesting for the same reason that you mentioned, which is like the volume of data is just so much. And which meant like, you know, for me, it's like a kid in a candy store where you could actually like, you know, build a lot of really cool applications with like all the volume of data that's like nobody else has, right?
2:48The other part that was quite interesting was just the physical world operations aspect of it. So maybe to answer your question, Going back, so Samsara for people who are not familiar with is a business where we try to build solutions or software and hardware solutions for our customers who operate in the physical world. When you say the physical world, you can think about an example would be people that are working in transportation or people working in construction or field services, even the plumber who shows up at your door. Like when you have issues, you probably work with a plumbing company who have a lot of plumbers working in an area who they're going to dispatch one person.
3:34So we build software for all of these type of solutions, right? And for all these companies, when you think about it, there's a common themed problem in the sense of they all care about safety as a theme, both of the people that work there and also of like the assets or like, you know, vehicles that they operate, which usually can run in the thousands. They all care about efficiency. Like how do I operate really efficiently? And a lot of these businesses operate with very low margins. So this is very crucial for them. And then the world is becoming more and more sustainable today. So many of these companies think about sustainability as a goal and like, how can they move in that direction.
4:19So that's effectively the type of solutions that we try to offer in terms of with this data. Now, maybe I'll give you a couple of examples and we can even talk a little bit more of maybe deep dive a couple of them. But when you think about safety, and we cater to people who are driving in the real world, we actually sell sensors, which in this case could be a physical camera that you can install in your vehicle, which has AI running on the edge. in this sensor where it's able to give you quick feedback around driving behaviors. Like for instance, like, you know, you might be driving out on the road for a long time.
4:58It can detect drowsiness and actually tell you like, you know, Hey, like you should pull over for instance.
5:04Jon Krohn:That's like a camera in the cab of a truck or something like watching the driver or how to, how do you tell if a driver's tired? So it's actually like the camera is actually in the truck. So the thing that we sell is actually like, you know, safety cameras. So we actually have the technology where you can install the safety camera in the cab of the truck. And it is actually like, you know, watching and learning from the behaviors of people's like, you know, behavior as they're driving. And we run AI models on the edge, like obviously vision models that actually can detect for these behaviors. It basically does like vision modeling and then also some degree of temporal reasoning on the edge.
5:47And a part of the fun problem for us is like, you know, how do you run these models with the constraints of what comes together when you're running it in like, you know, constrained hardware? Yeah. That usually happens on the edge.
5:59Jon Krohn:That sounds really interesting. So if you don't mind me digging into this use case a little bit more, obviously you can't tell us things that are your proprietary secret sauce. but to the extent that you can can you tell us a bit more about this edge compute you know like what kinds of compute do you use on the edge to be able to do like it sounds like you'd be doing machine vision processing which can be pretty pretty complex pretty computationally expensive yeah so we actually um do a like at a high level we basically do a lot of our training in our backend for these models. It's primarily inference on the edge.
6:36So for training, we have our own machine learning platform, which we built in-house. It's a Kubernetes stack. And we also leverage Ray. And it supports training, deployment, and feature management, etc. And once we do that, we actually have a system that can automatically manage the various versions of models that we have and can deploy onto the edge and the inference happens in the edge so obviously like these models a lot of them are like proprietary that we build especially based on use cases we are also starting to experiment like with everybody like in terms of tiny models from like you know off the shelf tiny LLMs etc like for doing like voice translation etc on the edge as well so that's that's something new that we are doing but when you really think about a lot of the vision models, it's like proprietary models that we build internally, like, and it caters to different use cases that we offer, which primarily like in the mobile camera was like around safety.
7:38Jon Krohn:Right, really cool. Yeah, there's a lot of different ways that you could be setting up your machine vision algorithms to be specific to your use case, to be compute efficient on the edge there. And then so I know that, you know, you're not doing much training necessarily on the edge of course uh it's mostly inference on the edge but even you know that's not something that i have any experience with you know having cameras doing machine vision in real time is that something that's pretty lightweight and easy today is there a lot of hardware that can do that kind of inference in real time it's not very expensive or rare it depends upon like how you balance the constraints right so i think for us really um there is hardware that you You have to then operate with the hardware limitations, especially when it comes down to the inference models.
8:24That's usually what we try to do. It's not like we put in very expensive hardware out on the road, because we also have to think about the overall bomb cost of the hardware that we actually leverage. It's a balance, right? But ultimately, at the end of the day, what leads our decision making is more around what type of use cases that we want to operate with and what is the compute that we need and how do we do it smartly. I think a lot of the edge computational reasons are more along the use case itself. When you think about it, like, you know, if I have to do like an alert round predicting whether you're using a mobile phone, which is a common thing everybody's guilty of.
9:06So that is a use case that we need to be always operating on the edge, primarily because you have latency considerations and potentially connectivity considerations, etc.
9:18Jon Krohn:For sure. I can imagine when people are driving, they can be driving a truck anywhere and you're not always going to have cell phone signal access. Yeah, that's totally true. We actually, it's funny, actually, one of the customers I met recently, they were in this business of actually laying out electric poles. Like, think about, like, they're basically like operating, it's a utilities business. Their job is to bring electricity to places where they don't have electricity. So, which means they also are operating in an environment where they don't have roads. so we do hear that quite a lot and a lot of our customers because of the nature of customers we have like you know you you're operating in like these environments which are uh most definitely like you know connectivity is not a guarantee so we by default build with that intent that like you know uh connectivity is not a guarantee and how do you also build these solutions from a lens of can they gracefully degrade uh and like you know operate in that environment like where it could be off the grid for, let's say, like a week and come back.
10:22And that's fairly common as well.
10:24Jon Krohn:Yeah, I was basically only thinking about, you know, intermittent for a few hours loss of access to a server, a remote server. But yeah, you could potentially be a week or more in that scenario you just described where somebody's setting up the electrical infrastructure. There's no roads, no cell phone towers. That is a really interesting use case to have to be building for. And I guess a lot of your solutions, you need to be ready for that kind of situation. Cool. Well, so yeah, so we've gone into a bit of detail now. I kind of dug deep on this specific video camera one. What are other kinds of applications that you work on at Samsara?
11:01Yeah. The other one you could think about is like efficiency. And like when you think efficiency, there are two class of problems that we could maybe talk about. One is like, you know, fuel efficiency. Like fuel is usually one of the biggest spends for people who have like a lot of vehicles. Like I consistently meet customers who spend north of like$100 million a year just on fuel. That's like a very common thing. So for them, they would talk about like fuel savings in many different dimensions. And like a common one is on like fuel efficiency savings by finding which of your drivers are actually like not driving really well.
11:43Are there like mechanical issues on your vehicle? Which one should I prioritize? And how do I drive better fuel efficiency, which could quickly ladder up to like millions of dollars of savings for people, right? And the other use case on the space of fuel, which we've started recently working on, it's an interesting one, is on fuel theft. We actually realized that, I think there was a study that I read recently where it was like$130 billion problem in the world where fuel is actually stolen or like siphoned off vehicles, et cetera.
12:18Jon Krohn:Like fuel -$130 billion annually, I assume, being stolen from vehicles. I had not thought of that leakage. Yeah, I had not either. Actually, like it was quite interesting. So what happened was we'd heard about this from some specific customers in specific regions where like, you know, the cost of labor was very low and fuel cost is sort of similar globally, right? So there we actually saw more of a spike of this behavior. So one of the challenges, all our customers in that space were telling us, like who operate in that region was like, hey, could you build a good solution for this? And a lot of the solutions out in the field required, you have to go buy external sensors, additional sensors, and actually install them into your vehicles, right?
13:07So the challenge for us really was like, hey, how do you scale this solution for all types, classes of vehicles and that can operate globally? Because the problem that you inevitably encounter is the sensors can be very noisy and the variants can be different based upon the vehicle you're driving. Right. So there's no one single easy solution here. And we originally explored this problem by doing some very simple or heuristic-based approaches, like rolling median as an example, to smooth out data, to find variants in fewer levels and look for scenarios. But those started immediately, they were diminishing returns, I would say, where the lossiness or the noise was just too much.
13:53And effectively, that was not a problem that we could solve. So we inevitably had to move down the path of, okay, let's actually look at different techniques here in terms of how do we apply machine learning algorithms. Like actually we do more like a statistical model here where, which can actually smooth out the variance and like effectively like improve the hit rate when we find like, you know, the anomalies in fuel. So interestingly, like we originally did it for that one region. And then we decided, okay, we have all this data and customers globally. Let's actually scale this up and see what's happening.
14:33And what we realized is it's not a very regional problem. It's a global problem. So that was quite interesting for us as a learning, just taking the data and like trying to apply it for like a broader variance of customers.
14:46Jon Krohn:Data scientists, it's time to talk about your tech. with Windows 10 support coming to an end, now is the perfect moment to rethink your setup. Enter Dell AI PCs powered by Intel Core Ultra processors. These devices are built for the demands of modern data science, delivering faster performance, smoother multitasking, and the power to handle even the most complex workflows. Whether you're training machine learning models or analyzing massive data sets, these PCs are designed to keep you ahead of the curve. Don't let outdated tech slow you down. Visit dell.com slash shoppcs to explore how you can upgrade your device and elevate your work.
15:24That's dell.com slash SHOPPCS.
15:29Jon Krohn:So basically you came up with some kind of model, like something like you don't have to tell me exactly, but I'm imagining it in my head as like a logistic regression model to detect some signal, like to have some probability of this kind of adverse event happening. So you have all of these factors feeding into the model and you were able to get kind of labeled training data in one region. And then when you apply that globally, you found that we're seeing this all over the world. Yeah. The main part of the problem is the variance in terms of sensor readings. So think about this as the part of the problem is we could have like, you know, let's say you drive like a Toyota Corolla and I drive, let's say, like a like, you know, Mazda or something.
16:15Jon Krohn:Yeah, I'm driving a Corolla and you're driving a Ferrari. Okay, it's okay. That'll be great. But the general idea is these sensors are not equal, right? So when fuel level is reported by these sensors, there's a lot of variance and there's a lot of noise. You also have scenarios like when you try to think about this from a lens of big trucks, they usually have dual tanks. like meaning there's not just like one tank. The sensor reading is only from one tank and there's actually like fuel flowing between these two tanks, right? So the fundamental crux of this problem is about how do you, for you to do really high quality precision for like these detections, you have to have ways of eliminating the noise.
17:06And so we primarily do a lot more like statistical models, I would say, like in, and like, you know, tune from there, Like that's the approach we use. Yeah.
17:16Jon Krohn:Really cool. Fascinating application areas. And it gives us a good sense of how Samsara is applying AI across so many physical problems. We have, you know, you mentioned construction, transportation. We also have from our research, we pulled out warehousing, manufacturing, retail, logistics, and some public sector applications as well. So it's a huge, vast problem surface area. And the idea with Samsara is to be helping your customers operate smarter across all of those verticals with an open platform that's obviously built to scale, dealing with the trillions of data points that your company is regularly processing.
17:59Jon Krohn:So, yeah, so really exciting place to work, I imagine. So let's talk a little bit more about the edge compute. I found that pretty interesting. And so when you're doing this kind of edge computing and you're helping, say, commercial drivers operate more effectively, you know, digging a bit more maybe into that kind of cab camera example. How do you decide when you're going to be doing on device compute on the edge versus sending something back for processing on a server remotely? because I do understand, you know, in the previous example that we were talking about, it sounded like we were kind of doing all edge compute, but based on our research, it sounds like sometimes there are some things that need to go back and be processed on a remote server.
18:46Jon Krohn:And so like, how do you handle these trade-offs between latency and reliability and quality? Yeah, I think it's really comes down for us in terms of, you know, the constraints or constraints for the product problem, right? Like when you really think about the problem of like keeping people safe in the real world, like you want to be able to like ensure it's operating in like a reliable environment without like latency constraints or like connectivity constraints. And so mostly whatever we do in safety, a lot of those solutions usually operate directly off the edge, right? But we do a lot of these problems which are like don't have those constraints.
19:28An example would be all the fuel solutions that I mentioned. We have all of them operating directly in the cloud. There's no latency constraints in any of those. We also work on a lot of problems around routing, for example, plan creation. It's less of a machine learning problem, more of an operations research problem. you could say like where you're like running to trying to basically do like a traveling sales plan and like generate routes for your like you know customers in terms of what's the most efficient way for them to fulfill so all of that like happens more on the more like in the cloud the one other thing we are doing a lot like most other companies which I didn't mention is like you know we're starting to leverage LLMs it's becoming very easy like we have like a AI gateway like pretty much everybody does now.
20:23And like, you know, now that kind of opens it up where any product engineer could just like directly start building or like building with these capabilities, especially like common solutions like summarization, data extraction, all of that, like into the product. So we have a cloud dashboard, which is increasingly becoming more AI first, not from a lens of AI first for AI's sake, but really like how do you eliminate work when you see what people need to do and where could you get that leverage, right? So we're moving in that direction quite a bit. Like how do you understand the use case so we can actually like pull in more of like automations and like agentic AI with LLMs.
21:05Jon Krohn:Right, yeah, let's talk about that a bit more. I mean, you said how having an AI gateway, most firms are doing that, but actually that specific term, AI gateway, I'm not really familiar with that. So yeah, tell us a bit more about this. It sounds like it's kind of, Is this a way for engineers, scientists to have access to kind of like maybe pre-vetted models that you feel are safe? Is that kind of the idea behind the gateway? That is correct. I think it's, think about it as there are like multiple different models that we leverage and for multiple LLMs available today. Now, there could be like, you know, like the best use case for different, sorry, the different use cases catered to different usage patterns of LLMs in itself.
21:49Now, how do you have like one single gateway, which you can like a central endpoint, which is available within like the Samsara ecosystem, whereas a product engineer could actually like leverage that endpoint. and based on their specific use case, define the constraints to say, hey, this is what I'm trying to do and leverage this particular model for me rather than for building their entirely new facade or think about the layer or the facade for every individual model that exists out there. So we use OpenAI or the Anthropic models or even Gemini. And there's also like, we kind of built this gateway with a little bit of a layer of intelligence where you could be smart about things like cost.
22:40And also make sure we're keeping up with our own internal rigor around being enterprise ready in terms of security, compliance, whatnot. So all of that goes away from a data scientist or engineer's responsibility. And they could actually delegate that out to another team. And they just get it out of the box. So that's like a platform capability that we have. Nice.
23:03Jon Krohn:So the gateway kind of allows you to, it allows your engineers, your scientists, your developers to be able to take into account capability versus cost considerations right out of the box. That's correct. Yeah. And we also design, like it's also gives you feedback loops where like, you know, for you to, as you are like, almost think about as a training ground where before I'm ready to deploy this across all the customers and like, you know, get it really into production. You could even get like very quick feedback loops from like, you know, not just like try it out, but also like understand cost profiles, but also like understand like quality and promote to production.
23:40So it kind of gives you like this ecosystem, which you could use like an example solution that like, you know, which I thought was pretty cool that somebody built was like a product engineer in the org, like built this solution where like one of the common customer problems that we had heard was, hey, like we want to identify when our drivers are misusing the vehicle assigned to them. meaning lots of companies have policies when they assign you a vehicle you're supposed to use it for specific needs now the only way for you to do that historically was like you have to see through all the content of video footage or like randomly pick stuff and like or core sample for like individual right now you can technically delegate that problem to like a video llm model and then which can automatically, like with the constraints you offer, it can actually like flag things.
Read the full transcript
24:36So we actually built that, like a product engineer who's not done a lot of data science in the past was just able to pick it up and build it. Now that's the capabilities that it's unlocking. So which means like a lot of engineers, the lines have blurred quite a bit in terms of like, you know, these different engineering stereotypes that used to exist. Yeah. Yeah.
24:56Jon Krohn:Such a great example where, yeah, you're basically, you're saying that historically you might have needed an AI expert, somebody with a PhD in machine vision to come into your company and be able to have this algorithm that could automatically label your vision training data for you. But now any engineer in the business and probably even non-technical people, they could vibe code a solution in a lot of cases. Yeah. To say, I'd love to be able to label these data. this is what and and you could you could i can imagine a scenario where you could vibe code that that kind of solution very quickly uh yeah really really cool time to be working on ai for sure and uh yeah i'm sure we have a lot of listeners out there who have spent uh you know many years developing their chops on ai but i'm sure we also have listeners out there that are just you know able to power through now and so they can listen to an episode like this and hear ideas about like Like, oh, yeah, that's so easy.
25:55Jon Krohn:I can be coming up with these ideas and applying them at scale, even though I'm relatively new to the space. A question that I have for you that this conversation is leaning into perfectly is with the huge amounts of data that you have, 20 trillion data points. And I'm sure that is growing exponentially every year with 20 trillion data points at your disposal and these kinds of exponential advances in AI, whether it's, you know, this ability to vibe. code, this ability to have an algorithm that can automatically annotate vision data for you. How do you define your physical AI roadmap? What kinds of internal decision filters help you help executives or help people who report to you distinguish between big long-term platform bets versus something that is an impressive demo but might not have great operational payoff?
26:57So I think it's a good question. In fact, actually, we have a lot of that. I think at Samsara, I would say a lot of engineers generally, we run hackathons and so on. So there's lots of demo culture in terms of building things, which we think is great in a way. How do you filter for the noise when you think about it? So the main point in any company or any ecosystem still that holds true is it's not about the technology. It's really about getting the outcomes and driving impact for your customers. So one of the things that we always try to drive our roadmap around is what is the business problem or what's the problem that the customer is actually facing?
27:48And how does the solution that we offer create like either operational leverage for the customer or effectively is it moving the needle from a lens of like, if you put this as a product, will there be a willingness to pay? Like it's usually a very good filter, I would say, like, you know, whether you're chasing randomized ideas or like something that's actually concretely solving pain. So those usually are good filters. So we always started it from a lens of a customer and that's how the company operates. We rely on this notion, which like, you know, is very ingrained in the business called customer feedback loop.
28:27Like anything we build needs to be like shared with like a few customers and we get some feedback and then we trade from there. I generally think about artificial intelligence or like any of these new technologies as like incredible from a lens of like, you know, doing and trying new things. But ultimately, all the companies that win out there are the ones who put the customer at the heart of the problem, right? Like, and trying to give them something which actually creates leverage. So, and that's the approach we try to take as a filter whenever we look at problems to invest back in.
29:02Jon Krohn:I like that. So it's the idea of this loop all the way through to the customer to ensure that the kinds of projects that your team, that the engineering team at Samsara are investing their time in, are likely to have a good payoff to be valuable to your customers downstream. Cool. So yeah, so we've talked about now LLMs and the cool capabilities that all of us have access to now. Now, something else, another emerging technology that you've talked about in a blog post we caught is talking about quantum computing. So I wanted to get into that a little bit. In an article you noted, and we'll have a link to this article, it's in something called Diginomica.
29:43Jon Krohn:And so we'll link to that so people can read it in full. But you note that connected operations platforms already do many things that transport reimagines. and that quantum computing may continue to enable this by crunching, quote, unimaginable amounts of data. So you talk about something that would today take a traditional compute platform, 10 septillion years can be done in minutes with quantum. So yeah, tell us about what quantum might be able to do in your space around routing, predictive maintenance, scheduling. And yeah, go ahead. Yeah. I think for, um, when you really think about it, right?
30:28Like, uh, a very practical example is like routing and scheduling, you could say, right? I give you an example, which is like pretty much everyone who studied computer science has gone through like the traveling salesman problem, right? It's an NP hard problem, which means like, you know, it just scales exponentially in terms of like a computational needs for you to solve it. and anybody who's trying to solve that problem effectively uses optimization techniques right so like in a simple sense to say it's like a traveling salesman problem where you're basically saying for routing like a vehicle has to visit 20 stops find me the most optimal way for it to visit the 20 stops it's i do believe it takes there are a little over i think a lot more than trillion i think maybe quintessential, I think, is the next one-ten trillion number of combinations that exist.
31:20So I think over time, algorithms, we've gotten good at approximation algorithms and then reducing the search space of that solution space and then coming up with answers. But I think getting to optimal solutions for the space of route planning is a great opportunity with quantum. Now, we just talked about traveling salesmen in terms of the one individual route problem. Now, move that along to think about, hey, like I have to make 10 ,000 stops. I have 15 vehicles. Find me the most optimal. Yes, you can do that. It's also a very classic, you know, engineering problem called like the vehicle routing problem.
32:03And again, solutions today uses approximations, not like precision answers, right? Now, let's extend that two steps further. Now, we've talked about stops. Now, between stops, what happens is routes. Think about it from a lens of, I need to go from stop A to stop B. I need to parse a bunch of segments to get there, and I have options within those segments. Now, the problem again multiplied to become something bigger. Let's add one more variance into it. Like every time, whenever like we all drive, we know that like we actually have real time traffic lights. That's real time traffic, very real thing.
32:43Right. Now. Lots of what we are doing is like these approximation models are like just taking some predictive examples and then like not really using closer to real time data. And then like they're just trying to model, OK, this is what I think likely will happen. And then, right. So with technologies, like, you know, we, and compute sort of becomes really easily accessible and like the limits goes off your ability to do things with much more high precision, uh, is possible. Right. Like obviously with things like routing, um, you know, you can get to like really high precise and like, you know, a lot more, uh, efficiency, uh, the gains can be like achieved.
33:24there's always the argument we can make of like you know what is good enough which i think is uh like you know there's like a point of diminishing returns at some point but like the possibilities are endless and like the different variables you can react to are pretty much uh endless right so i think the what i generally think will happen is like today a lot of the technologies we have like help you assist make assisted decisions and i think that's kind of great for you to get into like a world where truly self-operating systems, I don't think we are anywhere close, right? Because like with these type of problems or like the compute capabilities, plus like, you know, the level of access to data to do make real-time decisions are like limiting factors in that scenario.
34:12Jon Krohn:Right. Yeah. Fascinating. It's going to be interesting to see all the application areas that quantum ends up making a big difference in. We've had in the past year, a couple of times, we've had Dell's global chief AI officer, John Rose, on the show. And quantum is one of the spaces that he's most excited about. He kind of, he makes an analogy toward the way that, you know, you've been able to see for decades that super powerful AI capabilities were coming around now. And that now, you know, you can see the same kind of thing coming years from now with quantum and it's super exciting yeah it's definitely like um it's yeah it's like the pace of innovation in technology is just incredible and like like super excited to even see like what it'd be like in the next couple of years so yeah yeah exactly really exciting indeed um so one other application area that is exciting for me that we haven't talked about yet uh is when i was when we were preparing the research for your episode, your team sent to me this video, which I'll put in the show notes on commercial navigation, which is an area that had never occurred to me is so important.
35:23Jon Krohn:And so let me explain this. So when I need to go somewhere, typically I'll use Google maps. There are other obviously navigation apps out there for kind of regular users like myself. So, you know, Apple Maps or, you know, some people might use Waze. You know, I think a lot of professional drivers, you know, Uber drivers are using apps like Waze. But all of those kinds of solutions don't take into account commercial considerations. So for example, in this video that I'll include in the show notes, right off the bat, it shows how there's this big problem in trucking globally where trucks will be directed to go under a bridge that is too low for them to go under and the truck gets destroyed.
36:13Jon Krohn:And it has a really, you know, dramatic footage is shown in this video. And so it sounds like Samsara has come up with a solution, which is a commercial navigation tool. So a tool that is useful for avoiding bridges that are too low and there's probably other kinds of scenarios where a commercial navigation tool is essential as well it's actually interesting that you picked on the low bridge uh problem right so so we actually learned about that problem through a customer like a customer in uk actually came out and told us like hey we actually get these uh issues a lot where like you know one of our trucks hits a low bridge they have a lot of them in the uk particularly and uh you know like uh it gets into like this process where they now have to pay a lot of fine like they also have to like you know it's almost like a strike against in their record like there's a lot of implications so we didn't know about that uh until like you know this customer told us and they said like you know can you do something about this so we did two things actually like the on commercial navigation which is like you know one of the things is like you know we we were starting to work on a product where like we've heard this from customers through and through that like you know hey the navigation that i have which is works great for passenger vehicles is not something that i can use for commercial needs and constantly the manifestation of that came in the direction of like most of our customers would end up like paying fines they would get up into situations like exactly what you had talked about in terms of low bridge scenarios, et cetera, and they wanted to make sure they're compliant.
37:54I think a lot of, for them, the driving need is also compliance and safety of their drivers, right? So that's a big reason. So we built this product called commercial navigation, which is exactly built for that purpose. So one of the key differentiations here is the navigation is aware of your particular context, which means what type of vehicle are you driving and then it can change the mapping context to suit essentially what you are driving. So if you are actually driving a truck with like a, let's say like, you know, a 20 plus foot trailer, you will automatically be sent down paths which are like, you know, compliant to where you can drive rather than like down roads because there's just a path that exists, right?
38:41So which is an important piece that like, you know, we build. And then we also built this capability where every organization usually has like different policies. Some people have actually told us like, you know, hey, like driver cannot make a right-hand turn ever. Like, you know, so that's like a no-go, right? So we've also built in capabilities where can you like, you can customize. Do you mean left-hand turn?
39:02Jon Krohn:They can't make a left-hand turn ever. Oh, I think left-hand turn, maybe. Yes. Yeah, yeah, yeah. Because that's the like. I think it was like, well. Actually, yeah, left-hand turn. Oh, it's in the UK. But it's in the UK. It's in the UK, right? So then it is right-hand turn. I think you might be right. Yeah. But I've heard this from primarily companies that operate in more of the domains of school bus operations, for example. So that's one that I'd heard about this in. And then the mobility services, which is where they are paratransit type services. So we actually have some customers in those domains who actually ask for specific...
39:43You could only do this one side turn based on region geography. And then we also want to enforce things like you always have to arrive at the right side of the road. So one of the unique things that we've tried to do is take foundational mapping capabilities, apply the vehicle context and individual constraints that are needed for our customers to build this end-to-end experience that's super compliant, take safety at the heart of it, And that's what we've tried to build.
40:15Jon Krohn:Really cool. I love hearing these use cases. And kind of to take a lot of the examples that you've provided, the use cases, Simsar use cases over this episode, it seems like kind of the general idea here is to be taking something that previously might've been manual and making it automated, making it algorithmic. And so, you know, things like things that previously might've been done in a spreadsheet are now done algorithmically and automatically. And so things like fuel costs, traffic patterns, bridge height can all be taken into account automatically. What are the kinds of organizational bottlenecks or hidden costs that you've seen when companies try to swap out human heuristics for algorithmic planning?
41:05Well, that's a good question. I actually came across one very recently, just like a couple of months ago, right? So I know I described this problem around the traveling sales plan plus the VRP for road planning. So we actually built this product on road planning. Let's let our customers... The goal we went out with is we have to get the most optimal routes for our customers. And a problem that gets an engineer excited. Okay, this is like an engineering algorithmic problem. Let's actually build something great. Now, we built that product and then we got it out in front of some of our customers.
41:40and many of these organizations have a role called as like a route planner who's actually like responsible to make sure these routes are made and ready to go for like dispatch the next day for the various routes that the person like that their fleet has to go deliver to right so for as a part of our like what i mentioned in terms of feedback loop with customers like we went and showed them like you know hey like this is these are the orders you need to deliver tomorrow we created the most optimal routes here you go and the person goes wait i can't recognize any of these routes and they actually said something that was quite interesting they were like we have this driver sam who's actually need to go do this delivery to this particular grocer and they expect them at like 10 a.m tomorrow morning right now that's like human context that the route planner always had and there was a lot more like human to human engagement reason for why they wanted it that way.
42:37So they wanted to preserve some of the relationships that they had between driver to the actual customer of theirs. And we just lost all of that because we were just thinking of this from, it's an engineering problem. Let's build the most optimal routes. So what we decided to do was build an iteration from there of this product where we realized, hey, we have to meet the customers where they are. And one of the challenges is going to be, it's not like everybody's looking for just complete pure automation on the other side to get to something that would ultimately be like replace the human, get the system to do everything.
43:18whereas what effectively people wanted was like in an assistive experience how do you get me operational leverage but then don't sort of destroy what i've already built here in terms of like cultural value of the human connection etc so what we ended up doing was uh again take took the team technology that we built but then adapted it to say like we'll actually instead of giving you magical routes. We'll actually take in your current routes and we will work to iterate from there. We'll give you suggestions based upon your current inefficiencies we identify. So we will bring you along in this journey in terms of the transformation.
43:58So you're getting to efficiency, but then we are also learning from you in terms of the human constraints that exist in these problems. So this is a very interesting learning for me, which also resonated where I was like, okay, I always wonder of this world where should we have road planners? What would we do to actually take pretty much most of the work and automate it? But then it gave me perspective to say what we really need to be building today is assistive intelligence, which people can leverage because there's the human component, which is incredibly valuable for people in business and like running successful companies.
44:39Jon Krohn:Great example. I'm glad that you were able to dig into that. I thought it might be a really tricky question, but you were right there prepared with an excellent answer. And so speaking of preparation, you've had a huge amount of success in your career leading high growth engineering teams all over the world, uh chicago amsterdam berlin belgrade bangalore and so on and samsara seems like an amazing place to work you fast company uh listed it as one of their most innovative companies in the world uh the fortune future 50 ranked samsara as number seven uh in a recent uh ranking there and so that's like super exciting double digit percentage growth in terms of revenue year over year for many years so if people who are listening want to work with you want to work at samsara um yeah are you doing any hiring and what do you look for in the people that you do hire um we definitely are hiring um um as you mentioned the business is growing pretty fast and i think the main key part we generally look for uh are like you know we want people who actually show curiosity is one of the main things that we look for, like in terms of like really, this is a very fascinating and interesting space, right?
46:01So we think anybody who's a builder needs to really like get fascinated by the problem space in terms of what physical operations could mean, like want to meet customers and we enable that and like, you know, learn from how operations work. So I think that's a very important piece to like, you know, building great products, like making sure like the builders really understand the audience and the problems. So curiosity is like one of the key ingredients that we really look for. And then effectively, like, you know, it's going to be agility, I would say, like the other one, like where, this is a high growth startup, right?
46:38And like startups, like in general, like, you know, I would say like, it's a scale up right now. We could say like, you know, it's a public company, we're continuing to scale up. Like we still like to operate like a bit of like a high growth startup, which means like, you know, a lot of adaptability and agility is a part of it, right? So not everything's figured out, not everything's like super bureaucratic. Do you actually love an environment where you're enabled and you can move fast and show curiosity and actually like build with focus from there on and like get solutions out the door? That's usually what we look for.
47:12Jon Krohn:I love it. Moving the needle on projects, really making it happen. exactly it does sound like one of the key things um and praveen this has been a fascinating episode in general you're doing outstanding really exciting work uh how can people be following you after this episode to get more of your thoughts or to get more on what's going on at samsara you can actually follow me on linkedin i should be pretty easy to uh find i'm not super active on socials, I would say. So yeah. Too much work to do. Yeah, I think it does get hard when you're primarily operating every day. For sure. So LinkedIn's a good way to follow, keep in touch.
47:59I do often talk to people, especially people reach out in terms of interesting problems that we're working on and want to talk about exploring new ideas, solutions. So definitely welcome that. Do reach out.
48:15Jon Krohn:Awesome. And then I ask all of my guests for a book recommendation, and I already know what yours is. I can't wait to hear more about it. Yeah, I know we were just talking about it before we started. Like the book actually that I, one of the books I read during the holidays was this book called Do Androids Dream of Electric Sheep? It's actually like a fairly old book. I think it came out in the, I think, 50s or 60s. I don't remember. So this is actually a book that explores the idea of a world where androids and humans coexist. And what does it actually mean to be human versus being an android?
48:57It was a very fascinating read. I actually really enjoyed it. And it's actually very appropriate for the times we live with the base of velocity with respect to the changes that's happening. It's not a very heavy read. It's actually a story novel. Like, recently I've been reading more, like, novels, actually. But this one, I actually enjoyed it. And I thought, like, you know, something that everybody, this particular audience would enjoy it.
49:26Jon Krohn:Yeah, I love it. So it's, yeah, 1968, Do Androids Dream of Electric Sheep? It's by Philip K. Dick, a pretty well-known American science fiction writer. I haven't read it myself, but I've got to. And I'm just really quickly looking at a little bit of information on it online as you've been speaking. And I didn't know that the book is the basis for the film Blade Runner. Oh, yeah. Yeah. I think the, yes, I think it's been a long time since I watched Blade Runner. I think they take some concepts from Blade Runner, but I don't think reading the book is special. So I would say read the book. Yes, exactly.
50:05Jon Krohn:Yeah, and I think it's like, it's not necessarily the same plot. It's just kind of like an inspiration. But yeah, I didn't know that. Really cool. Thank you for that great recommendation, Praveen. And it's been so much fun having you on the show. We've learned so much about the kinds of problems that you're tackling at Samsara with AI. And yeah, hopefully we can get you on the show again in the not too distant future to get more updates on the exciting applications of AI that you're applying on such massive scale. Yeah, really enjoyed the conversation, John. Thank you.
51:03Jon Krohn:theft is a$130 billion annually global problem, and that Samsara uses statistical models to detect anomalies despite noisy and inconsistent sensor readings across different vehicle types. He described that an AI gateway serves as a centralized endpoint that lets any product engineer leverage multiple LLMs while automatically handling cost optimization, security, and compliance. He talked about how commercial navigation differs from consumer GPS, and how the traveling salesman problem with even a small number of stops has a mind-boggling number of combinations, but how quantum computing could soon overcome this to enable real-time precision routing that accounts for live traffic data.
51:43Jon Krohn:As always, you can get all the show notes, including the transcript for this episode, the video recording, any materials mentioned on the show, the URLs for Praveen's social media profiles, as well as my own, at superdatascience.com slash 967. All right, that's it. Thanks to everyone on the Super Data Science podcast team, podcast manager Sonja Breivich, media editor Mario Pombo, partnerships manager Natalie Zajski, researcher Serge Massis, writer Dr. Zara Karche, and our founder Kirill Aromenko. Thanks to all of them for producing a super episode for us today and for enabling that super team to create this free podcast for you.
52:14Jon Krohn:We are completely dependent on you and on our sponsors. So you can support this show by checking out our sponsors links, which are in the show notes. And if you'd ever like to sponsor an episode yourself, you can get the details on how by making your way to johnkrone.com slash podcast. Otherwise share, review, subscribe, but most importantly, just keep on tuning in. I'm so grateful to have you listening and hope I can continue to make episodes you love for years and years to come. Till next time, keep on rocking it out there. I'm looking forward to enjoying another round of the Super Data Science Podcast with you very soon.
From the publisher
VP of Engineering at Samsara Praveen Murugesan talks to Jon Krohn about processing 20 trillion data points covering 90 billion miles across private and public sectors, how the company helps truckers who operate long hours and travel for long stretches without cellphone signal, and who they’re looking to hire to help this physical AI pioneer keep on developing high-impact solutions for real-world problems. And, if you’re looking to work for the company, there’s no better time to apply, and you’ll want to listen to the end of the show to hear exactly what Praveen looks for in new hires.
This episode is brought to you by the Dell, by Intel, by Acceldata and by the ODSC, the Open Data Science Conference.
Additional materials: www.superdatascience.com/967
Interested in sponsoring a SuperDataScience Podcast episode? Email natalie@superdatascience.com for sponsorship information.
In this episode you will learn:
(01:01) The challenges of working with logistics data
(16:43) Operating Edge AI in logistics and construction sectors
(28:43) How quantum computing might redefine logistics
(40:09) The real cost of swapping human heuristics for algorithmic planning
(44:45) How to get a job at Samsara
