I am an AI Research Scientist with a decade of experience in AI research and three years in production software engineering. My background is mathematical (game theory, dynamical systems and multi-agent strategic interaction), and I apply it to auditing economically consequential LLM decisions, from labor markets to social-media recommender systems, including how feedback loops, strategic behaviour and multi-agent interactions can make AI systems fail at scale.
I developed inverse-optimization frameworks for inferring agents' preferences, established the first formal taxonomy of feedback loops and biases in ML systems, and designed scalable LLM evaluation and auditing frameworks, fine-tuning workflows and RAG architectures across models and platforms. I was also part of the team that won a gold medal in the 2024 Kaggle ARC Challenge, one of AI's hardest benchmarks for abstract reasoning and few-shot generalization.
- AI Evaluation
- AI Safety
- Automated Decision-Making
- Multi-Agent Systems
- Algorithmic Fidelity
- Algorithmic Fairness
- Recommender Systems
- Responsible AI
Published inPNAS · Nature Communications · Scientific Reports · npj Climate Action · COLM · EMNLP · ACM RecSys · ACM EAAMO · IEEE CDC · L4DC · ECAF
