Creator, Software Engineer
Current- Open-source RAG framework that streamlines deployment, testing, and iterating on RAG chatbots on your own infrastructure, lowering the barrier to entry for teams looking to query their own documents using AI
- Implemented configurable architecture for upstream chatbot components allowing multiple RAG pipelines to be served from a common API endpoint (Python, FastAPI, LlamaIndex, MongoDB)
- Designed an administrative dashboard UI allowing configuration, testing, and monitoring of user-facing chatbots (React-Typescript, Tailwind CSS, shadcn/ui)
- Utilized AWS infrastructure to host a SPA dashboard and RESTful API capable of supporting ~15k concurrent users (EC2, nginx, MongoDB, PostgreSQL)
- Automated infrastructure deployment using Amazon CDK, reducing provisioning steps from 50 to 5
- Identified blocking operations with virtual end-to-end load testing and implemented background processing to increase viable concurrent users by 1.8 times (Celery, Amazon SQS)