Ratish Oberoi

I’m an AI Systems Engineer and former CTO focused on building AI infrastructure—RAG pipelines, agentic runtimes, MCP tooling, and backend AI platforms. I turn complex requirements into deployable systems with validation loops and observability baked in.

As Chief Technical Officer at Finovia, I owned AI platform architecture, backend execution, and the technical roadmap for retrieval, automation, and internal AI tooling. I led engineering through a ₹1Cr pre-seed fundraising phase, and I directed RAG and inference workflows covering model routing, validation checks, backend APIs, observability, and deployment planning.

Across my projects, I’ve built an autonomous engineering runtime with retrieval-aware repository intelligence, AST evidence extraction, reranking, cited context assembly, and runtime judge loops. I’ve also developed repository-scale retrieval infrastructure and an MCP observability platform for tracing tool calls, context propagation, runtime events, and security boundaries.

My open-source contribution trajectory includes FastAPI, Qdrant client, LiteLLM, and OpenWebUI, where I’ve worked on runtime reliability, infrastructure improvements, and testing/validation practices. I’m motivated by building systems that are not just smart, but diagnosable, resilient, and easier for teams to operate.