Bangalore
We're building a proactive AI that manages your money the way family offices do for the ultra-rich — it knows your complete financial picture, watches it continuously, and tells you what to do next.
India has fewer than 1,000 registered investment advisors for 10 million+ investable households. Good financial advice never scaled — with AI, it finally can. But an AI advisor is only as good as the engineering behind it. That engineering is this role.
About the role: design the agent architecture — tool orchestration, retrieval, memory, and guardrails. Build the eval pipeline that decides whether financial advice is correct, and gate every change on it. Engineer for production: model routing, caching, cost and latency budgets. The LLM narrates; code calculates — you'll build both sides.
Who we're looking for: you've shipped LLM features to real users and have the war stories to prove it. Strong Python and backend fundamentals. You know structured outputs, tool calling, and RAG beyond tutorials. You think in evals, not vibes. Comfort with ambiguity, speed, and a roadmap that changes weekly.
We've previously worked at Groww, Peak XV (fka Sequoia India), JP Morgan, and Flipkart, and have raised a large seed round led by marquee VCs. We're currently in stealth — that's deliberate.
You'll work directly with the founders, before launch. Compensation: competitive salary + ESOPs. Bangalore, in-person.