San Francisco, California, United States
• Created the Hyperbolic AI Cloud VS Code extension in TypeScript to manage GPU instances; deploy Huggingface models via vLLM with gated auth, custom tokenizers and tensor parallelism; launch Jupyter kernels; and manage S3 buckets.
• Developed a TypeScript tRPC endpoint with PostgreSQL, and Amazon Web Services EC2, RDS, and VPC to compute supplier earnings and refunds, reducing billing reconciliation time by 70%.
• Built Hyperbolic CLI using Go, Redis, and Firebase with automated browser-based authentication to allow developers to manage compute resources from their terminal.
• Engineered a sub 100ms latency speech to speech AI Agent backed by a RAG knowledge base; deployed serverful Dockerized Python backend and communicated with frontend via Starlette websocket API.
• Led development of Hyperbolic AgentKit, a Python LangChain AI Agent framework for remote GPU compute orchestration; adopted by 100+ developers.