Members-Only
Recent Talks & Demos are for members only
You must be an AI Tinkerers active member to view these talks and demos.
ExpertFinder - Harvard E115 (Spring) Final Project
Demonstrating how to build an end‑to‑end expert search system using embeddings, RAG, ChromaDB, DVC versioning, and GKE deployment with CI/CD pipelines.
ExpertFinder is the final project for Harvard’s E115 Spring 2025 course (Advanced Practical Data Science) that helps users identify credible experts through a natural-language search system. It integrates natural language search with embeddings and Retrieval-Augmented Generation (RAG), stores profiles in a vector database (ChromaDB), and tracks data with DVC. The system calculates a credibility score using features like experience, education, and citations. It is deployed on Google Kubernetes Engine with containerization, Kubernetes orchestration, Nginx ingress, Ansible automation, and CI/CD via GitHub Actions.
RAG-powered expert finder leverages LLMs, LinkedIn/Scholar data, ChromaDB, and Kubernetes for deployment.
Medium is a web publishing platform using analytics, handling server errors.
Compose Email
Loading recent emails...