Arian Huapaya

Multi-modal AI pipeline that researches, scripts, narrates, scores and renders short videos. Hexagonal architecture (ports and adapters), Streamlit human-in-the-loop UI, Docker

LangGraph agent answering natural-language questions over a 20,000-row fitness dataset: Gemini router, ChromaDB RAG, scikit-learn model, MLflow evaluation, Streamlit UI.

Power vs precision grip detection from egocentric video using MediaPipe, YOLOv8, MiDaS and a small MLP

No project description available.

A python package to download weather data

A dexma API client implementation