Dibya Deepta Mishra · Chicago
I use economic models, causal inference, and machine learning to understand demand, evaluate interventions, and study incentives. My experience includes pricing research at Amazon, economic consulting at Cornerstone Research, and research on AI decision-making.
PhD in Economics, Rice University · MS in Computer Science, Georgia Tech
How should a retailer evaluate discounts and bundle prices? I combined product-text embeddings and gradient boosting with demand modeling, reducing forecast RMSE by 5% relative to standard logit models. I presented pricing recommendations to inform future A/B tests.
Do subsidized menstrual products increase adoption? We use doubly robust difference-in-differences to estimate a 9.6 percentage-point increase in pad use. The interactive pharmacy planner illustrates efficiency and equity tradeoffs in choosing locations.
Can a model encode the better option without choosing it? I use hidden-state probes and activation tests to study the gap between information represented inside language models and their choices.
In a separate Apart Research sprint report on agent collusion, I contributed to experiment design, Python labeling using Modal, and writing.
Estimated demand and causal price effects in competition and litigation work, and managed 2–3 analysts through code review and reproducible analysis.
Combined demand modeling and machine learning to forecast product demand and evaluate discounts and bundle prices.
Ongoing discrete-choice demand research on willingness to pay for AI capability and safety, and the economic value of capability gains.