Research engineer across Meta, AWS, and ASML, working on multimodal foundation models, post-training, and distributed training of 10B–1T+ models across thousands of GPUs.
How model width and depth shape training costs.
Jun 2025 - Jul 2026
Owned foundation model from 0 to 1: architecture research, large-scale distributed pretraining, and post-training (experiments across sample efficient data-mixtures, SFT, RL). Continually trained on billions of ad examples across image, text, and structured modalities.
Sep 2024 - Jun 2025
Designed pretraining scaling studies for 10B to 1T+ parameter LLMs across thousands of GPUs, improving parallelism strategy, compute-communication tradeoffs, cluster topology, and resolving NCCL bottlenecks
Nov 2022 - Aug 2024
Founding member of a site-wide AI project with Microsoft spanning ML modeling and enterprise LLM integrations. On the core team, built distributed data pipelines, error detection, and testing for EUV light sources.
Feb 2022 - Nov 2022
Trained deep learning computer vision models for detecting defects in semiconductor chips.
Sep 2022 - Dec 2023
Concentration in AI. Completed concurrently with full-time work at ASML.
Faculty-advised research applying deep reinforcement learning to news data, completed for research credit.
Oct 2020 - Dec 2021
Completed degree in ~1 year, graduating at 19.
Cum Laude and Highest Honors in the Major.
Brooks Lab (Jan 2021 - Dec 2021): A year of applying ML and software to computational genomics research, hired as a paid research assistant for the final quarter. Second author, PMC publication, 2022.
Tech4Good Lab (Jan 2021 - Mar 2021): Deep learning NLP research classifying coding for education.