Research profile

With my core training in Statistics and Biostatistics, my research focuses on the development of robust statistical methodology, with applications spanning precision oncology, high-dimensional multi-omic integration, and public health modeling.

For a complete bibliography, please refer to my Google Scholar profile, ResearchGate profile, or my CV.

💡 Note: Publications highlighted in teal represent first-authored works.

Filter by research topic (Methods selected by default). Publications are numbered chronologically, matching my CV.

Selected industrial projects applying core statistical methodology and machine learning models to large-scale ads auctions, platform economics, and predictive user behavior. Click on each project to expand descriptions.

📊 Large-Scale Auction Diagnostics & Bidding Optimization
  • Bidding Diagnostics Infrastructure: Engineered large-scale data processing pipelines to ingest, deduplicate, and analyze billions of daily auction logs. Developed custom database schemas to measure campaign budget pacing and bidding inefficiencies.
  • Bidding Recommendation Scaling: Authored technical requirements to optimize Target CPM (Cost-Per-Thousand) bidding recommendations for complex, flighted advertising campaigns. This resolved a critical system limitation, expanding recommendation coverage to unlock $10M+ in annualized advertiser spend and $100M+ in budget headroom.
  • Product Integration: Led cross-functional alignment with software engineering and UI teams to surface pipeline diagnostics as visual budget warnings directly in the advertiser-facing Google Ads console.
📈 Causal Revenue Modeling & Counterfactual Pricing Estimation
  • Causal Proxy Modeling: Designed a mathematically bounded causal proxy using short-term revenue signals to replace inflated advertiser value estimates, successfully correcting for selection and survival biases in platform-wide counterfactual pricing models.
  • Pricing Experimentation Operations: Assumed full technical setup, engineering operations, and statistical analysis for global subscription pricing experiments, implementing robust variance estimation and bootstrap workflows.
  • Statistical Decision Support: Owned the statistical narrative and presented revenue-foregone analyses to senior leadership. Conducted confidence interval width sensitivity analyses across 100+ global markets to secure alignment on platform pricing rollouts.
🔮 Predictive Online-to-Offline (O2O) Campaign Modeling
  • 0-to-1 Statistical Architecture: Led a greenfield machine learning initiative to predict campaign-level conversion rates (pCVR) for online-to-offline actions, establishing the data collection, tracking, and pipeline architecture from scratch.
  • Robust Multi-Stage Modeling: Designed multi-stage models combining relative outcome differences, predictive conversion components, and inverse-bid proxies to manage heavy-tailed outliers and zero-inflation in sparse campaign metrics.
  • Error Evaluation Metrics: Implemented customized error metrics to evaluate revenue models against large outliers, ensuring that large, high-volume campaigns do not skew model evaluation.

Upcoming Presentations

  • International Indian Statistical Association (IISA) Conference (December 26 - December 30, 2026 | Banaras Hindu University, Varanasi, India)