Abhimanyu Pallavi Sudhir

Formal education



Markets and AI (PhD work)

My current work focuses on building an algorithmic model of market dynamics to design AI agents with a market-based structure, and developing prediction market mechanisms to elicit beliefs about latent space variables to boost interpretability.

Related write-ups and talks.

Consistency checks and forecasting (Berkeley SPAR – spring 2024)

Ongoing collaboration with an ETH Zurich team supervised by Daniel Paleka to develop a consistency benchmark for LLM forecasters, as part of the Berkeley Supervised Program for Alignment Research.

General mathematics (Undergraduate work and prior)

Academic service

Courses and workshops attended

Other projects

Equivariant learning (2021-22)

Final-year MSci project with Professor Jeroen Lamb at Imperial College London exploring equivariant learning and causal DAGs.

Report: abhimanyu.io/legacy_writing/Imperial_reports/m4r.pdf

Lie theory (2019)

Undergraduate research project with Professor Richard Thomas at Imperial College London on Lie groups and algebras.

Report: abhimanyu.io/legacy_writing/Imperial_reports/urop.pdf

Presentation: abhimanyu.io/legacy_writing/Imperial_presentations/lie_theory.pdf

Lean (2018-19)

Computerized formal proving in Lean with Professor Kevin Buzzard at Imperial College London.

PhysicsOverflow (2014-15)

Co-founded PhysicsOverflow, a postgraduate-level physics Q&A site and open peer review system. See en.wikipedia.org/wiki/PhysicsOverflow for more details.