This AI agent automates the repetitive parts of data science How it works: upload a CSV → the agent scans column types, nulls, dupes, and outliers → LLM-planned cleaning (pandas executes it) → auto-generated charts → ML problem type + model suggestions → one summary report. The LLM plans and narrates; the data work stays deterministic. Stack: - FastAPI + pandas + matplotlib (backend) - Next.js + Tailwind (frontend) - OpenAI for planning and narration To build your own: run a Python API and a Next.js app side by side. Add your OpenAI key, start both servers, open the UI, upload a spreadsheet, and describe what you want in plain English (e.g. profile the data, clean missing values, suggest models to predict X). The pipeline runs in the background while you watch each stage complete. More agent projects coming soon! #ai #agent #tech