Building AI systems that help us understand, predict, and improve complex real-world systems
ML Intern @ AIBT & Regev Lab, Genentech, South San Francisco
2nd year PhD student @ AIMM, EPFL under Prof. Charlotte Bunne
Bachelor's + Master's @ IIT Bombay, Institute Silver Medalist (Rank #2)
Exchange student @ EPFL, Switzerland in Computer Science for Autumn 2022-23
Contact: eeshaan (dot) jain (at) epfl (dot) ch
[
Eeshaan Jain
] [
eeshaanjain
] [
eeshaan_jain
] [
eeshaanjain
]
I have a keen interest in applications of machine learning to life sciences, especially multi-scale representation learning and drug discovery along with geometric deep learning. Previously, I have worked on optimization, graph neural networks, graph retrieval, and fair learning under Prof. Abir De and Prof. Soumen Charkabarti at IIT Bombay. I have also worked with Google Research for a year, and with AWL, Inc. and Sony, Japan.
I have acted as a Teaching Assistant in the following courses:
- CS 461: Foundation models and generative AI (250+ students) [EPFL]
- CS 119: Information, calculation, communication (300+ students) [EPFL]
- CS 419M (Introduction to Machine Learning) (100+ students) [IIT Bombay]
- CS 768 (Learning with Graphs) (50+ students) [IIT Bombay]
- CS 769 (Optimization for Machine Learning: 100+ students) [IIT Bombay]
- MA 207 (Partial Differential Equations: 200+ students) [IIT Bombay]
- MA 108 (Ordinary Differential Equations: 400+ students) [IIT Bombay]
- CH 107 (Quantum Chemistry: 400+ students) [IIT Bombay]
I have received scholarships and grants from:
- Swiss AI Initiative (42000$)
- Broad School of MIT and Harvard for ICLR 2025 (3500$)
- NeurIPS Travel Grant for NeurIPS 2023 (2000$+)
- Google Research (Google Conference Scholarship) (2500$)
- IIT Bombay (Institute Academic Prize)
- Govt. of India (Institute Silver Medal)
- Heyning-Roelli Foundation (Semeseter Exchange Scholarship)(6000$+)
I have given the following invited talks:
- Agents and Benchmarks for Molecular Tumor Board @ AstraZeneca, Applied Data Science Group, Feb 2026
- Test-Time View Selection for Multi-Modal Decision Making @ MLGenX Workshop, ICLR 2025
I have reviewed for: NeurIPS 2025, ICLR 2026, ICML 2026, NeurIPS 2026
Highlights
VirTues is the first foundation model for spatial proteomics. Pretrained on 12,000+ multiplexed images from 5,300+ patients across 30+ cohorts and four imaging technologies, it runs cell segmentation and typing zero-shot in a single pass.
Nature Code
An agentic benchmark simulating Molecular Tumor Board decision-making through multimodal, longitudinal oncology questions, paired with a foundation-model tool framework that improves LLM reasoning and reliability.
Paper CodeNews
- Virtual Tissues (VirTues) — Our foundation model for spatial proteomics was published in Nature 🎉
- MTBBench — Our benchmark on multimodal sequential clinical decision-making (MTBBench: A Multimodal Sequential Clinical Decision-Making Benchmark in Oncology) was accepted to NeurIPS 2025 🎉
- AI-powered Virtual Tissues — Awarded Best Paper at ICLR 2025 MLGenX Workshop 🎉
- Test-Time View Selection for Multi-Modal Decision Making — Accepted as an Oral (top 4%) at MLGenX Workshop @ ICLR 2025 🎉
- Clique Number Estimation via Differentiable Adjacency Permutations — Accepted to ICLR 2025 🎉
- Co-organizing the Learning Meaningful Representations of Life (LMRL) Workshop @ ICLR 2025
- Two extended abstracts (Graph Edit Distance Evaluation Datasets, Graph Edit Distance with General Cost) accepted at LoG 2024
- Graph Edit Distance with General Costs Using Neural Set Divergence — Accepted to NeurIPS 2024 🎉
- Started my PhD in the AI in Molecular Medicine (AIMM) Lab at EPFL under Prof. Charlotte Bunne
- Best Poster Award at Eastern European ML School by Google DeepMind
- Winner — Kinase Selectivity Hackathon @ ML for Drug Discovery Summer School
- Efficient Data Subset Selection to Generalize Training Across Models — NeurIPS 2023 (with Google AI & UT Dallas)
- Research internship at Aalto University under Prof. Vikas Garg