Academic • AI/LLM/NLP
About The Project
Built a semantic assistant over crawled campus resources with chunking and ChromaDB vector retrieval. Implemented multi-turn dialogue, context carryover, and dynamic prompt routing for diverse GT queries using LangChain + ChatAnthropic. Created ingestion pipelines for structured/unstructured pages and prompt templates for reliable responses.
Achievements
End-to-end crawl → chunk → embed → retrieve → answer pipeline
Multi-turn context and router prompts for topic-aware query handling
Modular components for fast updates as campus content changes
More Projects
Links
sunny.sunho.park@gmail.com
