Honorable mention for the Jane Street Graduate Research Fellowship (2025).
Named an MLCommons Rising Star (2023).
Paper 'Pretrained Hybrids with MAD Skills' accepted to COLM 2025.
Paper 'MLGym: A New Framework and Benchmark for Advancing AI Research Agents' accepted to COLM 2025.
Presented 'Compute Optimal Scaling of Skills: Knowledge vs Reasoning' at ACL 2025, revealing distinct compute allocation optima for knowledge versus reasoning tasks.
Introduced Manticore: a method to construct pretrained hybrid models from non-hybrid bases without pretraining.
Developed Meta MLGym: a framework for benchmarking and developing AI research agents.
Published works on scaling laws, data mixture balancing (R&B), and reward models for code verification (including ICML 2025 workshops and preprints).