Bektaiyr Tuleshov

Owns the data and evaluation layer of AI-Expert, a system that detects errors in technical documentation and replaces a manual review process. Defines metrics for each release and scores against reviewer decisions, ensuring accuracy rests on agreement with human labels.

Built a computer-vision pipeline for microscopy images, replacing manual feature counting and artifact classification. Also built a generation pipeline for solvent discovery and contributed to perovskite solar-cell research, presenting at WCCE12 & APCChE 2025 and co-authoring at MRS Spring Meeting 2026.

Analyzed questionnaire response data for the WellMen diagnostics platform and rebuilt the intake into an adaptive flow. Removed repeated questions and shortened the user funnel while keeping the flow within medically certified question sets.

Replaced repeated numerical solver runs with trained surrogate models in Python, turning a slow batch workflow into near-instant results. Validated outputs against numerical benchmarks and rebuilt the training pipeline for stability across larger simulation domains.

Activities and societies: President's Scholarship; Capstone: ethylene plant design in Aspen Plus, designed acetylene hydrogenation unit and developed reaction kinetics.

Pursued a Bachelor of Engineering in Chemical & Materials Engineering, with coursework in numerical methods, transport phenomena, and process simulation. Awarded the President's Scholarship for research contribution as a Research Assistant.