ML Problem Design Suite – AI Reasoning Benchmark Dataset Authoring
I designed a library of computationally intensive STEM and ML problems for advanced reasoning and coding skill evaluation, with full documentation and validation. These datasets targeted AI benchmarking and model reasoning tasks, ensuring all prompts and solutions adhered to advanced reasoning requirements. My efforts supported internal and external benchmarking of AI model capabilities through comprehensive scenario coverage.• Authored 80+ ML/STEM problem-solution pairs for AI evaluation • Ensured dataset quality through rigorous prompt validation workflows • Applied Python stack for problem coding and testing • Produced all materials with high-standard technical documentation