Led full-stack development of GPT-4 features across backend (Python/Go) and ReactJS UI systems. Built the Function Calling framework with schema validation and sandboxed execution for AI agents. Developed RAG APIs using embeddings and vector indexing to improve grounding and reliability. Designed multimodal pipelines for vision, audio, and text with preprocessing and validation layers. Engineered low-latency real-time systems (WebRTC, streaming APIs) for scalable AI agents. Contributed to model evaluation workflows, including dataset creation and automated verification. Led architecture reviews and improved code quality and observability across systems.

Software Engineer at OpenAI

September 1, 2021 - June 1, 2023

Built a conversation orchestrator backend powering multi-turn reasoning, memory management, and context evaluation. Engineered real-time streaming APIs (SSE / WebSocket) delivering scalable and reliable model inference. Designed internal evaluation dashboards for benchmarking model outputs and analyzing error patterns. Implemented full-stack validation flows for verifying AI-generated responses and enforcing policy compliance. Diagnosed and resolved large-scale performance bottlenecks, reducing API latency from ~15s to ~2s. Developed internal debugging and code-review tooling.

Junior Software Engineer at OpenAI

June 1, 2019 - August 1, 2021

Built core GPT-3 API backend components (Python/Go REST services), including authentication, rate limiting, caching, and token streaming. Implemented batch and streaming optimizations that doubled throughput and reduced latency under production workloads. Created monitoring dashboards tracking latency, error rates, and performance benchmarks. Developed early AI output verification logic for detecting and blocking harmful or invalid responses. Built onboarding tools and reference implementations across Python, JavaScript, and Java SDKs.