• Building AI systems to automate regulatory and privacy compliance and continuous auditing across Meta’s Ads Delivery stack

• Building real-time and batch serving pipelines for multi-modal AI models, with xB+ daily inferences for privacy-aware Ads ranking.

• Scoped and delivered multi-year projects with extensive cross-functional collaboration. Key involvement in org-level roadmapping, interviewing and mentoring engineers

Leveraging computational intelligence to bridge the gap between design and manufacturing for Intel's leading technology nodes.

Designed and developed full-chip software tools for AI-based pre-silicon debug and defect repair of Intel’s latest CPU/Server products and test chips.

Developed numerical methods to study models of superconducting materials. The quantum mechanical nature of the particles and the strong interaction between them makes it an exponentially complex computational problem, intractable for realistically large systems. I formulated and developed approximate methods that circumvent this complexity in polynomial cost.

Mentored over 150 students and led weekly recitations in General Chemistry. Mentored (privately) two undergraduate students at City College, NY and Columbia, respectively, on General Chemistry.

‌• Developed classical molecular dynamics simulation to statistically model the self-diffusion in liquid para-hydrogen in regimes where hydrogen nuclei behaves quantum mechanically (ultra-low temperatures). Used Mathematica and Fortran.

‌• Reduced molecular dynamics simulation run-time by ~100x to calculate diffusion coefficient with results matching experimental observations.