Hey, I'm Shrey!

(Shreyans Sethi)

I’m a Computer Vision graduate student at Carnegie Mellon University and a UC Berkeley Computer Science alum.
I spent three years at IBM Research as a Software Engineer on a Biomedical AI team, working on multimodal foundation models and molecular generative modeling.
My goal is to build multimodal AI systems that improve human well-being and efficiency, with a strong focus on healthcare applications and 3D learning.

Google Scholar · LinkedIn · Email

IBM Research

Primary Software Engineer on a Small-Molecules AI team.

Open-Source Multimodal Foundation Model
Graph, image, and text modalities

Authored Paper on Model Use for Molecule-Target and Property Prediction
Accepted to Advanced Science

STAR-VAE: Latent Variable Transformers for Molecular Generation
Accepted to AAAI 2026 Workshop on AI for Drug Discovery

Internship Project

DPM360: End-to-End Model Training Framework for Healthcare AI

Featured CV & NLP Projects

Cross-Resolution Gaussian Splatting (XR-GS)
3D reconstruction for mixed-resolution input views with strong global structure and resolution-consistent fine detail.

Adapting Large Vision-Language Models for Medical Triage
Part of Team Chiron in the DARPA Triage Challenge, focusing on rapid triage in mass-casualty events.

Condescension Classifier for NYT Headlines
Annotated dataset, logistic regression model, and NLP analysis of linguistic markers of condescension.

Subfield Survey on Multimodal NLP
2022 survey summarizing architectural developments in VQA, VCR, and image captioning.

Relevant Coursework Projects

3D Computer Vision (CMU)

Computer Vision (UC Berkeley)