Projects
-
QuickQ: Built for a hackathon - a job search platform that uses AI to match candidates with roles and provides interview feedback. Backend uses Flask + MongoDB with Google Cloud Vertex AI, frontend is React + TypeScript + TailwindCSS. Learned a lot about integrating multiple AI services and building responsive UIs under time pressure. [Code Backend] | [Code Frontend]
-
MLX LLM Benchmark Tool: A benchmarking tool specifically for testing language models on Apple Silicon Macs. Handles memory management automatically, generates detailed HTML reports, and categorizes models by size. Useful if you're working with local LLMs on Mac and want to compare performance across different models. [Code]
-
SciQuery LLM: A research assistant that searches through arXiv papers using RAG (Retrieval-Augmented Generation). Uses Sentence-BERT for embeddings, FAISS for vector search, and DeepSeek for generation. Built with Gradio for the web interface. Good example of combining multiple AI components into a working research tool. [Code]