I've built production backend platforms at DeccanAI, including a human-in-the-loop reinforcement learning system coordinating 500+ annotators and supporting $10M ARR. I also architected Stark-Forge, a Kubernetes-based distributed backend framework with pluggable SQL, JSON, and Python verification services.
I own delivery of 20+ production MCP servers that generate verified execution traces. The Jira MCP server handles 10,000+ requests per day in evaluation pipelines used by Microsoft and Google, while my evaluation work improved end-to-end task success from 41% to 68%.
Previously at Deutsche Bank, I architected a RESTful ETL pipeline automating 10,000+ alerts daily and reduced data load time by 40%. I also built high-concurrency Java microservices that reduced outages by 30% and improved P99 latency from 820ms to 310ms for 5,000+ daily users.
My work spans Python, Java, FastAPI, Spring Boot, Kubernetes, data pipelines, LLM agents, and distributed systems. I've also built ML tooling at AiDash and shipped projects for financial workflows and large-scale product price tracking.
