Vinay Badhan

Leading automated code generation and testing platform for clinical trial study building through LLM integration and fine tuning. Implemented AI-powered code generation agents with custom RAG pipelines for auto-generating study extensions, tests, and deployments. Migrated from a custom agent engine to Google's ADK, improving reliability and reducing maintenance. Architected a self-serve platform enabling sponsors and CROs to independently build and validate studies. Transformed study setup from a ~6-month manual process to under a week with end-to-end automation. Collaborated with professional services to automate deterministic workflows. Reversed profit margins from -90% to +40% per study by removing bottlenecks. Implemented USDM JSON standard as the source data for the self-serve platform. Designed a rule-based evaluation engine for context-aware notifications. Load tested all services and scaled an India-based engineering team (6 developers).

Leading the development of an AI-driven study automation platform. Designed and implemented LLM-based code generation and RAG systems with versioned embeddings, async validation, and rollback-safe workflow generation, reducing manual customisations by 95%. Improved per-study profit margins through system automation and operational cost reduction. Led and mentored a team of 6 engineers; owned architecture, delivery, and production reliability.

Architected cloud-based IDE and assessment platforms supporting 10× traffic growth and global concurrency. Built a custom HTTP + WebSocket routing engine to optimize real-time code execution and collaboration. Reduced infrastructure costs by 55–60% via container optimization, autoscaling, and workload re-architecture. Led backend systems for candidate evaluation, scoring pipelines, and analytics at scale which provides skills insights for candidates and processes 20M records daily. Progressed from intern to Tech lead through consistent technical ownership and impact.