Projects

Case studies, not screenshots.

Every project here follows the same structure: problem → architecture → decisions & trade-offs → metrics → lessons learned. The parts that matter when a system has to survive production.

  1. Multi-agent architecture for measuring the energy consumption of LLMs based on prompting techniques and context length

    Prompt engineering optimizes for answer quality and ignores the energy bill. This multi-agent harness measured 2,700 LLM inferences to put a number on the trade-off.

    Python / LangGraph / Anthropic API / Ollama / CodeCarbon / EcoLogits / SQLite / GreenAI

  2. Certification exam simulator with an original question bank

    Certification prep leaves you a choice between ten official sample questions and leaked exam dumps. This generates an original, domain-weighted bank from the official guides and drills it with spaced repetition.

    Python / FastAPI / SQLite / Pydantic / React 19 / TypeScript / Vite / Gemini API / Anthropic API / Ollama