AI Writes the Code, But No One Understands It Anymore

The next time your favorite app crashes or your bank's website goes down, remember this: no one really understands what went wrong anymore.

AI systems are now writing most of the code powering our digital world—through tools like Cursor, Windsurf, and GitHub Copilot—creating software so complex that even the engineers who deploy it can't fully grasp how it all fits together. When things break, traditional troubleshooting becomes a nightmare of guesswork and blame-passing, with teams spending hours or days trying to find the needle in an ever-growing haystack.

Enter Anish Agarwal, the MIT-trained AI researcher who left academia to tackle what he calls "the biggest hidden risk of AI." His company Traversal just emerged from stealth with $48 million from Sequoia and Kleiner Perkins, building an AI site reliability engineer that can automatically diagnose and fix software failures. In this conversation, Anish reveals how they're already cutting resolution times by 40% for major cloud providers—and why the future of engineering depends on solving this invisible crisis.

Anish: My name is Anish. I'm the CEO and co-founder of Traversal. We came out with $48 million in seed and Series A funding, which were led by Sequoia and Kleiner Perkins. We've just announced our Series A raise and we came out of stealth.

I was fortunate enough to get into Columbia's faculty and I like thinking about theoretical problems, mathematical abstractions, and I think that university is an amazing place to do that. The thing that changed, and that was the time when everything with ChatGPT was happening. And so it just felt like something incredible has happened in the world, and it's like a once in a lifetime thing where the world has fundamentally changed. People don't even realize it. It just felt like this almost religious experience as to what was happening in the world.

I also saw some of the smartest people around me. They're either at OpenAI or Anthropic or Meta, or they're creating companies, and that's what makes me really excited. And so I think starting a company felt to me like a great expression of that.

Anish: I like sports. I like competing a lot, and I always naturally gravitated to math and science. I just liked how abstract and clean it was and it felt quite universal in what you could do. I came to MIT just because I thought machine learning and AI was really important and I wanted to understand it really deeply. This is like 2016, so about 8-9 years ago.

Once I got there, I think that one of the biggest moments in my academic research career was NeurIPS 2017, and the keynote was given by the Google AlphaGo team, and I just found it incredible that a system can learn this creative thing by itself because the complaint you always had before was that it's just copying people, but this was purely learning creativity by self-play. So I was like we should apply that everywhere, this kind of architecture.

I really like uncertainty. I like creating something from 0 to 1. Similar between research and entrepreneurship is that the uncertainty—you have no idea what's happening most of the time, and you have to find ways of creating structure from nothing. I think obviously the difference is in here, the time spans are compressed. In research, you get 5 years, 10 years to make an impact. Here you get one month. So the feedback cycle is very quick. But I think in this age of AI when AI is changing so quickly, being in that quick feedback cycle is actually very important.

Anish: We're building an AI site reliability engineer. So what that means is when you have large complex software systems and they break, we troubleshoot to help you figure out why it broke and then help fix it automatically.

Fundamentally what we're doing is when we have large complex software systems, and they break, we help figure out what happened, having a team of on-call engineers 24/7 looking at all of your data. So by the time an engineer comes onto a Slack channel, the root cause is already given. So rather than them spending so much time trying to figure out what happened or calling more and more teams—because typically what happens is you'll have one team look at the data, they're like, 'oh, it's not my fault,' then they'll call another team and another team and another team. That's how we go from 5 people to like 80 people on a channel. And so then rather than 50 people over an hour, it's like 5 or 10 people for a few minutes just verifying the answer.

The cost of downtime annually for all enterprises is around $400 billion so it's a huge problem. One of the biggest problems is downtime, is troubleshooting.

Anish: It's sort of like Perplexity. When you ask Perplexity a question, it kind of gives you evidence, right? It gives you citations of how it got to that answer. In our world, the citations aren't web links, the citations are links to your observability system.

Now we can publicly talk about—we've been working with them for the last 6 months now very closely—Digital Ocean. So they're a large public cloud service provider. I think they're the 3rd largest actually by number of people using them as their cloud provider. I think they have over 600,000 people using them as their core infrastructure, and you can imagine when they break, every person that is using them feels the pain because that's the main thing powering their system.

We've been working with them for 6 months now and we've found that in that 6 month period we've dropped the time to resolution by over 40%, like 37% to be exact, which is incredible, because as I said, every minute of downtime is worth thousands if not millions of dollars.

Anish: When we started the company in January of 2024, we started without an idea, but we had a clear taste of the type of problem we wanted to take on. So we wanted to do something that was at the intersection of our research, which was in causal machine learning and reinforcement learning, and how it intersected with AI agents.

Causal machine learning is the study of cause and effect, and what you want to understand is how do you get these AI systems to pick up cause and effect relationships from data. An A/B test is an example of learning cause and effect relationships. A clinical trial is another example. So these are like basic ways of running an experiment. And so that's what the study of causal machine learning is, and we're trying to see how that intersected with AI agents.

We went through a few different ideas. The person who joined us, our fourth co-founder Ammon, pitched us the problem of dealing with incidents. And as we looked into it, it kind of felt like a perfect problem—finding this needle in the haystack with many fake needles everywhere. So it fit with our research in causal machine learning and reinforcement learning really well. It fit with LLMs really well because the haystack is composed of logs and metrics and traces and code and configuration files and so on. It fits with AI agents really well because you have to automate this complex workflow where you're querying all these different pieces of the software, you're reasoning over them and then you're writing more queries and it's like this sequential adaptive flow.

And it's a big market because everyone cares about software not going down, and I think it's only going to get bigger because so much more code is being written now by companies like Cursor or Windsurf or Copilot or what have you. Just like Cambrian explosion of code being written. No one understands it in some ways. And so when software breaks, it's gonna be really difficult to troubleshoot.

Anish: Creating an MVP is so easy with all the tools out there. You can really iterate quickly with putting a product in people's hands. We built our first MVP probably last year in June or July, like about 3 months in, and with small companies, it worked great because the scale of the data was small, we could kind of look at their historical incidents, see what the playbook was, and then put that into an AI agent. And so our accuracy was like 90%, something amazing. So we felt really confident that this is gonna work.

Just because it works one time doesn't mean it's always gonna work because the world is constantly changing. And then we hit some of the larger enterprises, including Digital Ocean. Our accuracy went to 0%, which is very difficult to see as a tough week. Creating an MVP is easy, but creating a production system that works in complex environments is really hard. And so I think once you're not confusing an MVP with a production AI system, those are two very, very, very different things.

But then we rearchitected a lot of things. We said, how do we make sure that we are no longer trying to use our creativity and get that into an agent, but really use what these AI systems are good at, which is using computation, using inference. That's really what unlocked us and suddenly our accuracy went back up to 90%.

Anish: How do you make sure that your system gets better with the reasoning models? Because there are a lot of people we saw in competing companies where as the reasoning models came out, they didn't get any better. So how do you make sure that you're exploiting what the reasoning models are good at?

And the way I put it is that the reasoning models are very good at detective stories. You're a mystery novel and you're trying to figure out who did the crime, and there's all these different pieces of evidence that you're seeing. You're trying to figure out who is the person who did it. Connecting all those dots and figuring out the person who did it, that kind of detective story type workflow, I think these reasoning models are very good at, where you have a clear answer at the end and you have lots of moving pieces that you have to connect the dots between to get to the clear answer.

That felt kind of perfect for us because for us, you have all these different symptoms that are happening at the same time, and you need to find ways to connect the dots to find that specific right answer, like who did it. Making sure we were exploiting them to the maximum was crucial, I think.

Anish: AI is gonna write so much more code and no one really understands all of it. If you wrote all of it, you have in your head just how it all fits. And as you get to bigger and bigger systems already, you work with some of the largest Fortune 100 companies, no one has full context. No team has full context about what's happening because it's such a complex system. And that's just happening not just at the large companies, but also at the small companies because the code is no longer being written by a human, but by an AI system. So the lack of context means that when an incident happens, it's just so much harder to debug it because you just don't have all of the context you need.

And actually it's already happening—most people now are not developing code, they're starting to validate QA code or troubleshoot code, and that doesn't scale. That's the big problem. And I think the second big problem is that with all the developments happening with AI software engineering, our belief is that as engineers we get to do the really creative, fun work, architecting, system design. Over time, all engineers will be doing will be troubleshooting, which would be sad in my opinion. They should be doing the most creative work. And to make that a reality, you need to have systems, not just developing your software and building it, but also maintaining it. And so I think all our software maintenance needs to be reinvented.

Anish: You have to just kind of persevere. If something goes wrong, that's OK. I think in some ways it's a good thing because if it was easy, then everyone could do it. I think this is one of those problems where the problem statement is very easy to state, but to actually solve it is really hard. And I think that's where there's like a lot of companies trying it, but very few actually succeeding. And I think having the ability to stay resilient and have grit when something doesn't work and just stick with the problem is a big part of what differentiates us as a company.

Think about what's going to be important 10 years from now regardless of whatever happens in the world. You have no idea what's happening most of the time, and you have to find ways of creating structure from nothing. And so one way I say it is that in typically hard jobs, whether it's in finance or technology as an engineer, you have a point A, you get to point B, and it's very hard to get from point A to point B. In the world of research and also in the world of entrepreneurship, you don't really know where point A is. You don't know where you are, you don't know where point B is, you don't know where you want to go. And if you did know, it's still very hard. And so you're constantly in this game of trying to just decide where you are and where you're trying to go.

Anish: What's interesting in this job is that the stresses can be very high. Lows can be very low, so I think getting used to just very high highs and low lows is important. But I think the one thing I've learned is that the time it takes me to recover is very fast. Even if I'm super stressed, I'm super tired, within 1 or 2 days, if I just take it off, I'm back to full force. And so I think that's been like a good learning is that if you're really enjoying what you're doing, and even though there's these moments of massive stress, levels of stress that you'll ever face otherwise, if you really love a problem and you're attracted by it.

That's where it attracts other people. We all love the problem. We're all passionate about it and so we really feel like a great deep desire to solve it. And probably the most important thing out of anything is surround yourself with people that you care about, that you like, that you want to be like. And if you do that, life will be OK. I think about my research life, I found the right PhD advisor or advisors that really molded me and guided me in the right way. If I think about this world, I found the right investors that guided me at the start, the right customers that helped define the product. You live and die by the people you surround yourselves with and the people will kind of guide you through these different parts.

And I think building a taste for the right people to mentor you and be a partner with you is the most important thing. The rest of it, I think will figure itself out if you can find the right people around you. But I wouldn't just start a company for the sake of it. I think you should feel some deep desire because it's not easy. And so you need that kind of deep belief or motivation to sustain you over time.