For the past few years, I led the teams that shipped Deep Research, which can spend hours investigating nuanced questions across hundreds of sources. We built Canvas, turning natural language into professional-grade code. We integrated Gemini directly into Chrome, making AI assistance as native as bookmarks. Each release crossed another frontier. We were turning science fiction into reality.
But here's what I realized: as compelling as agentic AI is for consumers, agentic AI for enterprise is what remakes the economy. Agentic AI's full potential can be realized when it augments entire professions, changing how work gets done across industries. That's where you get GDP-level impact. I wanted to focus my AI experience on solving those kinds of problems. So I started thinking about the hard challenges where enterprises can benefit from agentic AI’s transformational impact.
That’s when I noticed something unusual. A startup called Resolve AI was recruiting more people from my DeepMind team than any big tech company or research lab has managed to do. People I'd worked with for years, all choosing this startup over other attractive opportunities. That caught my attention. When somebody is that good at recruiting the best AI talent, I thought I should at least meet them for a coffee and see what's going on.
Look, when you're at DeepMind, you get approached with a lot of interesting opportunities. So for me to actually leave, it had to be something special.
When I met Resolve AI's founders, I found something rare. Spiros and Mayank didn't stumble onto production systems as an interesting AI application. They actually lived the problem. They co-created OpenTelemetry (a ubiquitous foundation) to observe software systems. Now they're on a mission to simplify how we navigate and manage our production systems with Agentic AI.
The problem they are solving for is universal and imminent: Every company runs on software. When production systems break, everything stops. Revenue stops. This was exactly the kind of problem I was looking to solve with Agentic AI.
For such a young company, Resolve AI already has meaningful traction, real customers, and a clear product-market fit in a segment that naturally needs what they're building. It seems early, but in many ways Resolve AI has made serious progress where most startups are still betting on potential. They're delivering tangible value already.
It's rare to find the combination of a great team, deep domain expertise, early traction with customers, and the size of the problem they're tackling. It is a generational company in the making.
When I look at the problems AI could solve in enterprise, “understanding and debugging production systems” stands out as uniquely important. It's not just another process. It's THE meta-process that determines whether all other processes can function. Every company runs on software. When that software breaks, everything stops.
This is also a problem where Agentic AI's unique strengths align with the domain's needs.
Here's what we're building: Agentic AI for software engineering teams that helps them understand and navigate production systems. That's the north star.
Every new service or integration makes production systems harder to understand and manage. We're solving for the exponentially growing complexity of production systems.
Why? We want to amplify every engineer’s leverage by freeing up their time from understanding and debugging production systems. We want them to build and operate their production systems without barriers or dependencies.
If this problem sounds interesting and challenging for you, we’re looking to expand our engineering team. Check out the open roles here.
I'm joining the Change Agents Series on July 17, 2025 hosted by Corinne from Greylock to discuss the state of art agentic AI systems and how we envision the future of software engineering. Join our waitlist to be part of this conversation.
Stay current on how the best engineering teams are using AI in production. Customer spotlights, product updates, how-tos, and more delivered monthly.
Discover why most AI approaches like LLMs or individual AI agents fail in complex production environments and how multi-agent systems enable truly AI-native engineering. Learn the architectural patterns from our Stanford presentation that help engineering teams shift from AI-assisted to AI-native workflows.
Hear AI strategies and approaches from engineering leaders at FinServ companies including Affirm, MSCI, and SoFi.
Resolve AI has launched with a $35M Seed round to automate software operations for engineers using agentic AI, reducing mean time to resolve incidents by 5x, and allowing engineers to focus on innovation by handling operational tasks autonomously.
