Popular repositories Loading

1. file-tree-triage file-tree-triage Public

   Drive deduplication that produces a plan and never deletes anything

   Python 1

2. hardware-parts-data-pipeline hardware-parts-data-pipeline Public

   Scrapes product data from two robotics and electronics retailers and normalizes their differently-labeled specs into one typed, comparable catalog of 380 parts with converted units.

   Python

3. ai-product-extractor ai-product-extractor Public

   Turns messy product listings into a clean, validated, de-duplicated dataset with a per-row confidence score, routing anything uncertain to a human review queue before it reaches the database.

   Python

4. wnba-player-vs-defense wnba-player-vs-defense Public

   Turns a WNBA player's game log into an info sheet that joins every game to the opponent defense faced that night and splits production against tough versus soft defenses, exported to Excel and CSV.

   Python

5. invoice-to-structured invoice-to-structured Public

   Turns invoices, purchase orders, and bank statements into validated JSON and CSV, routing only trustworthy rows for auto-posting and sending anything suspicious to a human review queue.

   Python

6. rag-grounded-qa rag-grounded-qa Public

   Answers questions over a document corpus with a RAG pipeline that cites its sources, refuses when retrieval turns up nothing confident, and scores recall@k, refusal accuracy, and grounded rate.

   Python