CEO of Pathway, Zuzanna Stamirowska, joins Ryan Chacon on the IoT For All Podcast to discuss the current problems of supply chain logistics and how it is improving. She starts by introducing her extensive and interesting background and what Pathway is working on. She then goes on to talk about the impact of IoT in the supply chain and handling costs for companies looking to adopt it. Ryan and Zuzanna wrap up the podcast with a high-level discussion about the evolution of challenges and advice for choosing the right technology for your business.

About Zuzanna

Zuzanna is the CEO of Pathway, a game-changing Dev Tool for enterprise clients, enabling real-time machine learning on live event streams. Pathway develops a Database that Thinks™, which provides a unified source of truth from raw data in logistics and other industries. Zuzanna is the author of the state-of-the-art model for forecasting maritime trade published by the American Academy of Sciences (single-author, pre-PhD). While working on this project, she saw the mismatch between the business needs, the reality of the data in logistics, and the current ML and AI methods - this was the spark to launch Pathway. During her career, she was responsible for sales of SaaS technology underlying Spoj.com. She holds a Master's degree in Economics and Public Policy from Sciences Po, Ecole Polytechnique, and ENSAE, as well as a Ph.D. in Complexity Science.

Interested in connecting with Zuzanna? Reach out on Linkedin!

About Pathway

Pathway develops a Database that Thinks. A game-changing Dev Tool for enterprise clients, enabling real-time machine learning on live event streams. Pathway provides a unified source of truth from raw data in logistics and other industries. Their product provides application developers with a capacity for real-time incremental in-memory transformation of complex event streams. Pathway is a female-led, deeptech developer tool company. It was started by top people in data processing and AI. Pathway's CTO, Jan Chorowski, previously worked at Google Brain, and most of the team members are top-competitive programmers. Their mission is to significantly impact the physical economy by bringing automation, scalability, and repeatability to industries that are having a hard time capitalizing on the advances in AI.

Key Questions and Topics from this Episode:

(01:29) Introduction to Zuzanna and Pathway

(07:18) Overview of Pathway

(09:00) Problems in supply chain logistics

(11:54) Impact of IoT

(15:18) Handling costs with customers

(18:01) How challenges have evolved

(21:33) Advice for choosing the right technology

Transcript:

- [Voice Over] You are listening to the IoT For All Media Network.

- [Ryan] Hello everyone, and welcome to another episode of the IoT For All Podcast, the number one publication and resource for the Internet of Things. I'm your host, Ryan Chacon. If you are watching this on YouTube, please like and subscribe to this video. If you're listening just on a podcast directory, please feel free to subscribe if you haven't already done so so you get the latest episodes as soon as they are out. On this episode, we have Zuzanna Stamirowska, the CEO of Pathway. They are a very fantastic company. They are a company that's developed a database that "thinks" in their own words, a game-changing dev tool for enterprise clients, enabling real-time machine learning on live event streams. Pathway, the company itself, provides a unified source of truth from raw data and logistics in other industries. Very, very fun kind of conversation that we have around supply chain, logistic problems, how to handle cost questions, overall value that IoT provides beyond tracking of just shipments. We talk about the challenges in the space, how companies can get started, how to prove out ROI in this space. And she's a fantastic guest, and I think you'll get a lot of value out of this episode. But before we get into it. Any of you out there are looking to enter the fast-growing and profitable IoT market but don't know where to start, check out our sponsor. Leverege. Leverege's IoT solutions development platform provides everything you need to create turnkey IoT products that you can whitelabel and resell under your own brand. To learn more, go to IoTChangesEverything.com. That's IoTChangesEverything.com. And without further ado, please enjoy this episode of the IoT For All Podcast. Welcome, Zuzanna, to the "IoT For All" podcast. Thanks for being here this week.

- [Zuzanna] Hi, Ryan. It's a pleasure to be here.

- [Ryan] Yeah, great to have you. Let's kick it off by having you give a quick introduction about yourself to our audience.

- [Zuzanna] Sure. So my name is Zuzanna. I'm the CEO of Pathway. So I'm actually based in France, so Paris here, and I'm Polish, like my origins are Polish, and I'm actually here from a banks background, I'm a bit of a politician, game theorist, and went to military engineering school, ended up doing a PhD on forecasting of maritime trade, developed the state of the art model for forecasting of maritime trade, which gave birth to Pathway.

- [Ryan] Wow. So yeah. Tell us a little bit more about the journey for Pathway to exist and to get here and kind of the journey you've been through. It seems like you've been through a lot of different, or you've gone through a lot of different kind of industries and different kind of positions and kind of paths, and how'd you get into tech and entrepreneurship?

- [Zuzanna] Yeah. Indeed, it was quite a journey. So in fact, I mean, even though these were different experiences, I'd say they were all building up to the same direction. So I mean, when I was young, I wanted to study economics and law at the same time. So I went to the one school, actually the one university that offered the sort of courses, and that was Sciences Po, and in France, this is the school for French presidents to maybe make it more intuitive. And I mean, I studied economics, I fell in love with economics and economics is much more hard science, it's harder science than your usual humanities, let's say.

- Sure.

- Then I was at Stockholm School of Economics where I fell in love with game theory. And at the same time, I was trained in management consulting. So it was a fun, fun, fun period in my life when it was like both business and math theory. In fact, I saw how the two match, you know, 'cause when you do management consulting, you think a lot about strategy and, in fact, game theory helps you a lot in even just, you know, designing organizations, et cetera. So then I applied for a master's at the Ecole Polytechnique, which is a French engineering school where like some of the most famous physicists in the world actually worked, like Poincaré, for people who know him. Yeah, so I studied here at Ecole Polytechnique, and I also had happened to have friends who were competitive programmers. And actually, I kept on running with this crowd of competitive programmers, like some of tops in the world because like one of our co-founders, in fact, created Spoj.com. It was the first platform for training of programming and the first technology that enabled people to run the code online and to evaluate code online. This grew to a community of millions of people. And somehow, I mean, you know, the founders and the best programmers became my friends. So I knew the folks in algorithmics and very strong programmers. Yeah, so I mean, being with them, I sometimes went to the conferences with them. So I learned somehow by immersion, you know, things about algorithms, computer science. I'm not a computer scientist myself, but I got to work a little bit in the field, even published a bit.

- [Ryan] That's great. Yeah. And then, in fact, I was looking for a PhD, and game theory, especially when you talk about more algorithmic aspects, has natural applications to transportation. And I did a PhD, and you know, my PhD, I had a pleasure and enormous fun to work on 30 years of daily movements of the entire world fleet, maritime fleet. So I had like the daily movements of ships to analyze, and it was an enormous, super powerful data set. And I spent a couple of years on this at the Institute of Complex Systems of Paris and Sorbonne, and I designed that model for forecasting of maritime trade. I wanted to try to understand, "Okay, I mean, we have the systems, we have these flows, but how does it move? Why does it move, and what happens if something bad, like, you know, if you have a disruption? Where will the traffic go? How will the network reorganize itself?" And specifically, I wanted to know what sort of information is necessary, is really necessary to be able to do the forecast. Yeah, and it was fun because I found that it was just like one-parameter model. So something that, like, the economists were maybe not so happy about, because usually you would like to put much more assumptions into this, and not really. So yeah. I mean, I'll say it was a journey. It was like, you know, bits from here and there. I think it was just mostly my interests that were guiding me. And it's all built up to me becoming the CEO because I got, you know, like these skills coming from different backgrounds, so it's easier for me to see and imagine different functions. So I manage to look over the tech, but also over the business functions. Yeah, and it was also the spark to see really the need. The need to change something.

- [Ryan] Very cool. Very cool. Well, thank you for that. I wanted to kind of pivot here for a second and talk about the company first, real quick. If you could just give a quick, high-level overview of the company, just what Pathway is and what you do, the role you play in IoT. And then I'll dive into some other questions.

- [Zuzanna] Sure, so Pathway's an intelligent event stream database, but in fact, we call it "a database that thinks." So it's a developer tool. It's a developer tool, which allows you to very, maybe simply put, make sense of the chaos. So in the case of IoT, you may well imagine that you have IoT sensors deployed, for example, on containers in transport. So we have IoT movement, and the raw data that you get speaks the language of measurements. So we will get the data table where you have entries such as location of a container and where it was at a given point in time. That means that the questions you may ask this data are usually constrained somehow to the data columns that you have, to the objects that you have in the table. So it's not very powerful. Because your interests, your like objects of interest are the individual containers. So it's hard to ask about processes. And what Pathway does is it automatically and in real-time translates its raw data into the data that speaks the language of the business and the language the business cares about, which is anomalies, processes, locations. And so this is like very high-level explanation of what's happening. And there is some cool tech involved, but maybe not for now. I'll keep it for later.

- [Ryan] No, you're fine. I did wanna dive into, I know we wanted talk a little bit about kind of supply chain logistics problems, kind of the space in general, and so I guess high-level it for us. Tell us a little bit more about kind of supply chain logistics as from a industry standpoint, what the current problems you all see in the market are, and then we'll go from there.

- [Zuzanna] Of course. So, well, first of all, logistics, it's not such a digitized industry. Sometimes for people who come from Google or from big tech companies, it's shocking. Like, the point to which logistics is not digitized can be very surprising. 'Cause I mean, in theory you would assume that when you have a big company and you in fact have like 10 companies in the world that control most of the maritime trade flows, that they should know basically where their containers are, like what's happening in their operations, they should know it. Well, the fact is they don't, and this is for, like, simple reasons. There are different levels of digitization around the globe, so we may have different processes in different ports around the world. Some people may still be filling things on Excels, not necessarily pushing data to external systems. Then company grew by acquisitions, so just internally, like, putting data in one place is an issue. But then, of course, so this is the situation before COVID. So it was already, let's say, not so great, but at least we could rely on business as usual and read the operational knowledge of people in logistics. But with COVID, with the disruptions of ports, like ports being shut down in, for example, well, China, crews being pulled out from ships, this created a lot of backlog, both of cargo that couldn't be charged on ships and couldn't be moved. And also, the prices of containers went up at least 10 fold. And then the disruptions somehow spread. So when you think about logistics, it's not just, you know, the movement of a given container from A to B, but it's a system. So it's like when you block the Suez Canal. I know. Ryan, do you remember that thing? Yeah, it was 10 billion direct cost to the world economy. When you block the Suez Canal, you will feel the consequences in Los Angeles, and you will have all the other ports impacted because it's like a butterfly effect. In fact, so whenever you have these disruptions, they add up, one to another. And the situation is getting worse. We need to find ways, in fact, to try to react quickly whenever they happen.

- [Ryan] Gotcha. And one of the questions I do have is as we talk about not just its industry itself, but IoT as it connects, what do you all see as the overall value IoT provides beyond just the tracking of shipments? Which when we're talking about supply chain and logistics, that's kind of where a lot of people's head goes to is being able to track shipments, which is great, but there's other value that IoT provides and love it if you kind of shed light on how you all think about that and the kind of more full-scale impact that IoT makes.

- [Zuzanna] Yeah, sure. So IoT, well, right now is being seen mostly as a tracking devices, so it's useful for monitoring whenever you have like a more added value cargo. And of course this is great, 'cause then whenever you run into issues with your insurance, you need to find who was responsible for damage or something like that, I mean, you can trace it back, or sometimes you can see where your cargo is waiting, right? Is being idle. So this is great, but then the value of IoT's not only linked to the transport, to the cargo in question as you just said, but in fact, you can get much more of it if you think of IoT as a data source, as a data source and a data sample from which, then, you can draw conclusions about the overall processes. To give you an illustration, in fact, we have a client who's putting IoT devices on empty containers. And actually, their main business case is not monitoring of the cargo inside of the containers, but it's monitoring the empty containers themselves. For example, to increase the fleet utilization and to understand how their processes look like. So I'd say the first question is really to understand how my transportation process looks like, 'cause people don't know. They honestly don't know. And I had clients who launched investigations in like far away corners in the world to understand, "Whoa, what's happening in my process? Why do I have transportation that was supposed to go on train, you know, all of a sudden traveling by truck?" And this poses risks to cargo and then some contractual constraints. So first step, it's very simple. It's understand the process. And for this, you don't need to equip every container. You just need a sample, and then you can map your process automatically. So in fact, you avoid steep costs of data integration. 'Cause the alternative would be to maybe go through, you know, one subcontractor to another subcontractor, try to get data from them, any sort of information what's happening. Whereas with IoT, you get this multimodal, at least at level of container, a multimodal visibility of what's happening. And then once you get this, you may spot a potential for optimization. So I had actually a CEO of one of rather big logistics companies who told me, like, "Zuzanna, listen," you know, "I feel like there is so much we can improve, but I really don't know where." "I feel that it's there, but nobody can tell me where it is." So that's it. It's very simple, you know? The first step.

- [Ryan] One of the things you mentioned was about not having to equip all the different... It'd be 100% equipped, kind of, from a full-scale standpoint. You can start out smaller. And then you also mentioned it helps... There's a cost element to it that people I think are hesitant to adopt because of. How do you kind of handle that cost question, and what kind of advice do you have for companies that out there looking to kind of get started down this path but may have hesitations?

- [Zuzanna] Yeah, so cost of course is important and as important is the selection of the right sensors, 'cause this is something that we've seen. And so for example, for La Poste, we helped them to reduce the cost of IoT deployment by 50%. And that was thanks to the fact they were able to select cheaper sensors that were easier to maintain because they were doing, in fact, a better job for their use case. I'd say that like the best way to go is really to start small, select the sensors, and see what you can really get out of it directly. So do it directly with analytics because just running an IoT pilot won't give you much. I had many clients, in fact, coming to us saying, "Oh, we tried something. We were not even able to read the data." This was something that we saw. Whereas, I mean, many companies, in fact, have data already. Sometimes there's IoT, you know, that they use for other purposes. But they're not using the data, so they're not looking at it. So first step would be to take your data and see what's in it. And then of course the business case can be different from business to business. I mean, in maritime shipping, they're kind of traditional, I mean repetitive, this would be linked to visibility, and idling containers, and CO2 emissions. Yeah, whereas for postal services, this can be linked to operation excellence. Like this is what we've seen. But the first step, really, and it's not expensive at all. It's just have a couple of sensors, not much. Put them on your assets, but have a smart installation. So I mean, you need to have a technology to fit the data in and to make sense out of it. And the feedback we got from La Poste was that, for them, it was a game changer, 'cause before, they were thinking of those assets, and right now, they think in terms of locations, routes, processes. And for them it has actually deep business implications.

- [Ryan] Sure. Yeah, well, that makes a lot of sense. Now, one of the other questions I wanted to ask you was outside of just this specific kind of vein of thinking that we've been talking about here. What are some of the other challenges that you've seen across the industry? One of them, I guess, and this kind of goes back to what we were just talking about, but making sense of raw data, proving ROI, things like that. How have you seen those challenges kind of evolve and how do you kind of get around them, and advice for companies to kind of overcome those challenges?

- [Zuzanna] Yeah. So, well, first of all, we've seen an evolution like before COVID, throughout COVID, and now with the disruptions just piling up. So I'd say that before COVID, there was much more hesitation about, well, digitization in general in the logistics industry, but also of course, everything that linked to IoT, I mean, it was seen a bit like a nice have. Whereas gradually, during COVID, visibility has been, at least visibility has been seen as something as a necessity, and the question right now is more like, "How should we deliver it and what should be our place on the market in delivering visibility to the end customers?" So this has grown to a big pain. Now, then the question of how to prove the ROI, it really depends whom you are talking to, 'cause the ROI is rather easy to measure for shippers. So shippers would be, like in the U.S., you would call them, rather, "importers." So people who really, they care about the cargo that you have inside the boxes and, you know, every day of delay cause them money and they actually have it calculated. So I talked to one company, they told me that for them, like, for just one construction site, I believe, it was like 5,000 euros per day of delay. So that was just per one site, right? So it can grow actually pretty fast. And then also you lose market share. So they have the calculations, and when you talk to supply chain people there, it's a bit easier. For the logistics companies, it may be a bit harder because it's more projected, it's more a revenue for them that they would get, like, possibly from having a differentiator for their client, but it's also an enormous opportunity for them to take a stronger stand on the market with their position because they already streamlined so much data. They have this power of being pretty central and having data across different operations, so for them it's a market opportunity, in fact. And the easiest way, so the advice I would give is like the easiest way to find an ROI is to try to link it to operational excellence, so any sort of cost optimization, 'cause then if you cast costs, you can clearly compare to an objective, and then you can calculate the ROI, even though we believe that the ultimate value for the logistics companies is really in the fact that, I mean, they have the data, they have the freedom, I mean the freedom and the position to monetize it.

- [Ryan] Yeah, absolutely. That's fantastic. Last question I have before we wrap up here is companies out there looking to get started are oftentimes put in a situation where they don't know how to choose the right technology, the best sensor, that kind of thing. What advice do you have for those companies kind of just looking to get started? We talked earlier about kind of the focus on data samples and then iterating from there when it comes to getting started, but when it comes to the actual, physical pieces that you would put together, how do you kind of advise companies on choosing the right pieces, and is that something that's oftentimes handled maybe by the company that they choose to work with that has a solution as opposed to figuring out themselves?

- [Zuzanna] From like, at least our clients, the one that we dealt with usually handled their sensors on their own since they were picking them and we were helping them to pick the best ones. But I wouldn't say that it was like a very painful or expensive exercise. A smart way to go about it was, first, not to trust the specs provided by the producers. So La Poste, for example, took a number of different sensors with different connectivity and they tested, they just put them like in the cheapest, in the easiest way, they just put them on their postal containers with different settings as well, and they tested which one works best simply. And it was pretty quick to decide which sensors made sense for them. And I'd say this is the best approach, to test the sensors also for the use case we have at hand. And then with Pathway, in fact, they were able to see directly on the map, I mean, they got the score for the quality and the coverage of given sensor. And indeed, it's interesting to have a look how different sensors work, because we've seen funny things. Like some producers may insert artificial data points in places, you know, to somehow try to seem like a superior solution. Sometimes you have sensors that are supposed to wake up with motion with accelerometer, but they don't wake up. And in fact, these are often more expensive sensors, so it may be cheaper to get the ones that will just send you signal now every 15 minutes, but at least you'll be safe.

- [Ryan] Yeah, for sure. Yeah. No, those are great insights for sure. I think it's always a challenge and an interesting conversation. Each customer and company kind of handled it their own way. Lots of options out there and trying to decipher between all that's very, very difficult at times. So, yeah. Overall, this conversation's been great. I really appreciate you taking the time. For our audience out there who wants to learn more about Pathway, about what you have going on, stay in touch, follow up with questions, what's the best way to do that?

- [Zuzanna] Yeah, so I mean, you can directly write to me at [email protected] and we'll be super happy to schedule a demo. And right now we're also looking for beta testers and any feedback we can get on our developer tool, because we'll launch it I think by the end of the year hopefully, so.

- [Ryan] Fantastic. Great. Well, thanks so much again for your time. Really appreciate it. And I'm sure we'll be doing more content together. I'm sure our audience is gonna get a lot of value outta this and anything we do in the future. So thanks again and look forward to hopefully having you back soon.

- [Zuzanna] Thank you so much, Ryan. It was such a pleasure. Thanks.

- [Ryan] Yeah, absolutely. Thank you. All right, everyone. Thanks again for watching that episode of the IoT For All Podcast. If you enjoyed the episode, please click the thumbs up button, subscribe to our channel, and be sure to hit the bell notification so you get the latest episodes as soon as they become available. Other than that, thanks again for watching, and we'll see you next time.