Themis Angelopoulos

Built a chat application with four AI-driven Greek philosophers on the user's own Claude or OpenAI key, including a Go engine, JavaScript screens, a Wails desktop window, and an Android shell. Implemented a safety layer with message judgement, reply checks, encrypted local storage, and logs without user text.

Built Moly, a writing coach available as a browser extension, desktop app, and Android app, with a Go backend and JavaScript screens. Integrated Claude and OpenAI with the user's own key, encrypted local storage, and crisis-language detection.

Built a Python platform (FastAPI, SQLAlchemy, Redis, Docker, pytest) that guided software projects through Socratic questioning, with a maturity score, workflow optimiser, and modular agent design. Published it as a family of Python packages on PyPI.

Researched trading strategies using quantitative data analysis and machine-learning meta-labeling techniques, following the methods of Marcos López de Prado and Alexander Elder. Tested every strategy against the lab's validation standards; none passed, which taught rigorous out-of-sample testing.

Built a modular trading system based on Alexander Elder's Triple Screen, with multi-timeframe analysis, machine-learning signal validation, and execution controls. Also built a Flask web application for an automated cryptocurrency trading bot.

Completed a certification in Machine Learning in Finance from QuantPy in 2023.

Completed a certification in Python in Finance from the University of Thessaly in 2022.

Studies in Regional Science and Economic Development at Panteion University.