Anish Agarwal
@_anish_agarwal
Founder @traversal_ai | Professor @Columbia | Searching for root cause
We all know how important inference-time compute is to AI. Making a bet on it last September (before it was cool) was key in developing an agentic architecture that allows @traversal_ai to work for complex incidents where no runbook exists, rather than just for simple alerts.…
At Traversal, we made an early bet on exploiting inference-time compute—and it’s paying off for enterprises. Listen as CEO @_anish_agarwal talks on @sequoia's “Training Data” podcast about how this decision is leading to significantly faster, more accurate incident resolutions.
As a former management consultant, I can't resist a 2x2 framework :) At Traversal, we put one together to help customers navigate the dynamic AI SRE landscape. Check it out here: bit.ly/4k6k7IU tldr: (i) most companies are only helping solve routine alerts; (ii)…
bit.ly/4k6k7IU ← Part 1 of our AI SRE blog series It seems like there’s a new “AI SRE” company every day. And of course, the incumbent observability players are all trying to build their version too. It’s no surprise — the problem is clear, the market is massive,…
Traversal solves the observability industry’s most daunting problem: fast, accurate, and reliable root cause analysis. But how? By combining the best of LLMs (reasoning models), agentic architectures (swarms of AI agents), and traditional machine learning (causal ML), we're able…
The AI SRE landscape is evolving quickly — and for enterprises, that’s a double-edged sword. The technology is badly needed, but it’s tough to know how to evaluate it. How do you measure accuracy? When should you build vs. buy? What will observability incumbents provide out of…
One month ago, Traversal came out of stealth to tackle one of the hardest problems facing businesses today: software reliability. We believe we’re the right team to solve it. Traversal combines deep academic roots with world-class product engineering. At our NYC office, you’ll…
Last week’s @RaiseSummit in Paris was fantastic. One key takeaway: Concern about troubleshooting and repairing code written by AI is growing. I had the opportunity to speak with engineers and execs from across the industry at the beautiful Carousel du Louvre, and it’s clear that…
Too many engineers spend holidays tethered to their laptops, dreading an incident. With Traversal’s AI SRE, you can get ahead of downtime, cut through noise, and resolve issues faster — before they take over your holiday. 🎆 Enjoy the long weekend. Let us keep watch. 🎆
Grateful to @Redpoint for the recognition - @traversal_ai was included in the 2025 InfraRed 100, which honors the most transformative companies in cloud infrastructure. And observability / SRE is of course ripe to be disrupted. The best part is that it happened while we were…
We were still in stealth when this announcement came out, so we couldn’t share it — but we’re proud to say that Traversal was named to @Redpoint’s InfraRed 100 earlier this month! Grateful to Redpoint for including us alongside others reimagining the future of tech…
AI SRE products have a lot of hype — and they’re all promising the same thing. Unless we figure out how to separate what’s real from what’s just good marketing, we’re headed straight for the trough of disillusionment. Yesterday, I shared our breakdown of the current AI SRE…
Why does the current state of site reliability feels like having two heart attacks a week? Why won’t AI-generated code make it into mission-critical systems? My cofounder, Raj, and I had a lovely time speaking with @BogieBalkansky and @sonyatweetybird on @sequoia's Training Data…
Grateful to @BryanOffutt for being a believer since day zero — and becoming a good friend along the way. Thank you for the support, perspective, and for encouraging our efforts to build great systems and a great culture. Honored to be in this with you and the @hanabicapital team.
@_anish_agarwal, Raj , Raaz, and Ahmed are building something special at @traversal_ai. I could write paragraphs about the underlying technology, the technical aptitude of the team, the nuances of the problem space and UX, how painful the problem they are solving is, or the…
It’s awesome when you feel instant alignment with someone, like we did with @charliecurnin on the need to rethink how engineers troubleshoot. Thanks for being such a thoughtful partner since day one. Here’s to many more years ahead.
believe it or not, we’ve been looking for a company like @traversal_ai for years at @sequoia for more than a decade, we’ve been obsessed with the “rise of the developer,” leading to partnerships with companies like @github, @MongoDB and @stripe 💡 more recently, it became…