I am proficient in Python, machine learning models, SQL, and database design, demonstrated through projects such as an ODI Win Predictor, a Football Analysis Tracking System (in progress), a Digital Learning Products Store Database, and analytical blogs published on Medium. My development workflow is powered heavily by Cursor AI (ChatGPT for coders), which I’ve used extensively for end-to-end coding, rapid iteration, testing, ML model training and evaluation, and data visualization—significantly accelerating execution and experimentation.
I’ve also leveraged NotebookLM to design graphical elements, mind maps, and interactive visuals, and n8n to automate repetitive, multi-step processes. Professionally, I’ve worked with Mercor, Alignerr, and Mindrift, contributing to the training, testing, and evaluation of large language models including ChatGPT, Gemini, Claude, and Google AIM. This included evaluating football-related analytical responses across the Premier League and La Liga using metrics such as helpfulness, severity, pattern loss, and factual accuracy.
In parallel, I worked on agent-style task automation projects, where I designed and optimized prompts, rubrics, instructions, and YAML configurations for AI systems handling real-world, daily-life tasks under strict success and failure thresholds. I’ve also contributed to AI red-teaming and adversarial testing, conducting ethical prompt-injection techniques (roleplay, instruction-following, double-persona) to probe model safety boundaries, along with multimedia annotation projects across video, image, text, and audio data. Additionally, I possess foundational knowledge of elite sports-analysis tools such as HUDL Sportscode, Datatrack, and Nacsport, strengthening my profile at the intersection of sports, data, and AI.