Software Development Engineer Intern
During my time at Podbook, I focused heavily on backend scalability and integrating AI pipelines into the core product.
The primary technical achievement was building a hybrid RAG pipeline for LLM-powered features and heavily optimizing the user onboarding flow. By decoupling notification services into an asynchronous Redis Queue pipeline, I brought the critical path latency down by 97%.
- ▹Architected a scalable backend using FastAPI and PostgreSQL, Built a RAG pipeline integrating BM25 and FAISS vector search with sentence-transformer embeddings, to improve context retrieval precision for LLM-powered features.
- ▹Created a non-blocking email pipeline via Redis Queue (RQ) to decouple notification services from the authentication flow, slashing user signup latency by 97% (from 4s to 100ms).