Internship • AI/LLM/NLP
About The Project
Developed a voice-enabled RAG chatbot with personalized retrieval over enterprise data (PDF/audio/web) using LangChain, FAISS, Whisper, and a three-stage semantic search workflow (ChatAnthropic). Evaluated and deployed ML models for lead scoring; implemented secure, modular back-end services (FastAPI, JWT, isolated vector stores) for product integration.
Achievements
+12% lead-scoring accuracy (XGBoost vs. baseline)
92% query accuracy for the RAG assistant
−35% response latency via embedding optimization & prompt chaining
+40% retrieval efficiency with tagging/cross-linking
Productionized a modular Python framework for the CRM product
More Projects
Links
sunny.sunho.park@gmail.com
