At VNGAG, I developed and improved internal energy trading systems using Python, Pandas, and SQL. I also built an automated pytest suite with 100% code coverage for core components.
I integrated external market data through Joule and Volue APIs, conducted research and data analysis for trading strategies, and migrated a standalone JavaScript trading application into the central Python bot infrastructure.
I built Vega, a single-user AI assistant with a CLI, a Textual terminal interface, and a local Starlette web service. Its main agent delegates to specialists using Claude, Codex, or a local model, including one in a network-isolated sandbox.
My projects also include an OpenAI API assistant restricted to an approved knowledge base, a local application that generates tailored CVs and cover letters as PDFs, and a Python pipeline that turns saved short videos into searchable Obsidian notes.
