Peer-reviewed work in computer vision (ICCV, IEEE IV) and recommender systems (ACM RecSys Challenge). Full list and citation counts on Google Scholar.
HINT: Learning Complete Human Neural Representations from Limited Viewpoints
Builds complete 3D human avatars from a handful of camera views using a sagittal-plane symmetry prior and explicit body-model supervision — designed for sparse-view setups like in-vehicle perception.
ScatterNeRF: Seeing Through Fog with Physically-Based Inverse Neural Rendering
A NeRF that disentangles fog from the underlying scene using physics-based scattering losses, enabling clearer reconstruction in adverse weather for autonomous driving.
United We Stand, Divided We Fall: Leveraging Ensembles of Recommenders to Compete with Budget Constrained Resources
Ensemble of lightweight recommenders that beats heavyweight competitors under tight compute budgets — entry to the ACM RecSys Challenge 2022.
Lightweight and Scalable Model for Tweet Engagements Predictions in a Resource-constrained Environment
Optimized LightGBM pipeline for predicting Twitter engagements at scale — finished 4th overall and 1st among academic teams in the ACM RecSys Challenge 2021.