Aug 2023 – Present · 2 yrs 10 mos · Bengaluru, Karnataka, India · On-site

• Developed a Cost Agent for production plants using FastMCP, and built the full frontend in Next.js + TypeScript with features like execution timeline visualization, typewriter controls, follow-up question handling, Developer Mode (tool calls, parameters, JSON logs), and support for adding custom MCP servers via UI.
• Developed an integrated multi-agent AI architecture using LLM orchestration, RAG (Retrieval-Augmented Generation), and context-aware memory to enable seamless collaboration between specialized agents.
• Implemented a judge agent pipeline that evaluated, ranked, and refined responses from other agents to ensure accuracy, coherence, and reliability before user delivery.
• Engineered high-throughput data pipelines in PySpark and Databricks, reducing infra cost by 50% and enabling real-time analytics on 15M+ vehicles.
• Developed and scaled a Generative AI chatbot for technicians to 2.5k daily active users.
• Built scalable web applications using React.js, optimizing diagnostic dashboards and UI performance by 35%.
• Designed and deployed backend APIs with Java, Spring Boot and Graph data structures for sales and charging systems.