🗺️ Breathless: My Research Internship Hunt
From Datadog to Hugging Face, a detailed chronicle of my 2025 internship hunt - doubts, strategy, failures, and four final offers.
💥 Building a Vulnerable Bank MCP - Then Automating an Agent to Hack It
How we designed a vulnerable banking app and used an LLM-powered agent to find and exploit its flaws automatically.
⚡ Saying Thank You to a LLM Isn't Free - Measuring the Energy Cost of Politeness
How even a simple "thank you" costs extra compute and energy when interacting with large language models.
⚽ Betting on Data: How to Build a Winning Football Prediction Model
Discover how we can use data science to predict football match outcomes and optimize betting strategies.
🤖 Teaching Poker to a Language Model: Fine-Tuning Mistral 7B - Introducing MistralBluff
Teaching Mistral 7B to play poker during the 2024 Mistral Fine-Tuning Hackathon.