Shitij Govil

about

Hi! I'm a third year Computer Science undergraduate student at Georgia Tech, where I've been fortunate to work with Prof. Pan Li and Prof. Animesh Garg. My recent works spans robotics (learning from unlabeled videos) and ML for science (currently particle physics).

I will be applying for PhDs for Fall 2027. My research goal is to create sample-efficient models grounded in real-world dynamics. I've recently become interested in robotics, and I have a few work-in-progress projects in this domain.

Outside of research, I love staying active and playing sports. I boulder and lift a lot, and occasionally play pickup ultimate frisbee and volleyball. I also listen to a lot of music (mostly techno) and DJ. Sometimes I also urbex.

Accepting opp applications.

research

Locality-Sensitive Hashing-Based Efficient Point Transformer for Charged Particle Reconstruction thumbnail

Locality-Sensitive Hashing-Based Efficient Point Transformer for Charged Particle Reconstruction

Machine Learning and the Physical Sciences Workshop, NeurIPS 2025

Shitij Govil, Jack P Rodgers, Yuan-Tang Chou, Siqi Miao, Amit Saha, Advaith Anand, Kilian Lieret, Gage DeZoort, Mia Liu, Javier Duarte, Pan Li, Shih-Chieh Hsu

Unified evaluation of HEPT and GNN baselines, and HEPTv2: a lightweight decoder that removes clustering and enables fast end-to-end inference for charged particle tracking.

Is Human-Written Data Enough? thumbnail

Is Human-Written Data Enough? The Challenge of Teaching Reasoning to LLMs Without RL or Distillation

AI for Math Workshop, ICML 2025

Wei Du, Branislav Kisacanin, George Armstrong, Shubham Toshniwal, Ivan Moshkov, Alexan Ayrapetyan, Sadegh Mahdavi, Dan Zhao, Shizhe Diao, Dragan Masulovic, Marius Stanean, Advaith Avadhanam, Max Wang, Ashmit Dutta, Shitij Govil, Sri Yanamandara, Mihir Tandon, Sriram Ananthakrishnan, Vedant Rathi, David Zhang, Joonseok Kang, Leon Luo, Titu Andreescu, Boris Ginsburg, Igor Gitman

Developed strong reasoning capabilities in base models without reinforcement learning via high-quality CoT.

Pre-training graph neural networks with structural fingerprints for materials discovery thumbnail

Pre-training graph neural networks with structural fingerprints for materials discovery

arXiv, 2025 [In Submission: Machine Learning Science and Technology]

Shuyi Jia, Shitij Govil, Manav Ramaprasad, Victor Fung

Pre-training GNNs for materials science using cheaply-computed structural fingerprints.

Using Hyperspatial LiDAR and Multispectral Imaging to Identify Coastal Wetlands thumbnail

Using Hyperspatial LiDAR and Multispectral Imaging to Identify Coastal Wetlands Using Gradient Boosting Methods

Remote Sensing, 2022

Shitij Govil, Aidan J. Lee, Aiden C. MacQueen, Narcisa G. Pricope, Asami Minei, Cuixian Chen

Developed a novel approach combining hyperspatial LiDAR and multispectral imaging data with gradient boosting methods to accurately identify and classify coastal wetlands.

projects

Efficient Skill-based Reinforcement Learning thumbnail

Efficient Skill-based Reinforcement Learning

Course Project CS 8803 DRL

A model-based RL framework that extracts reusable skills from rewardless offline data and reuses the data by relabeling with an optimistic reward estimator for efficient exploration through the learned dynamics model.

Winner-Take-All Sparse Autoencoders thumbnail

Winner-Take-All Sparse Autoencoders

Course Project CS 7461 DL

A winner-take-all sparsity mechanism for sparse autoencoders in the context of mechanistic interpretability. Investigates feature-wise sparsity constraints for SAEs to prevent dead features.