Azim Dehghani Amirabad

Azim Dehghani Amirabad

Principal AI Scientist at J&J, previously Postdoc @MIT, Ph.D. @Max Planck

About me

Hi, I'm Azim 👋

I am a Principal AI Scientist working at the intersection of AI, machine learning, and computational biology. I build foundation models and AI systems that integrate multimodal biological data, from genomics and proteomics to imaging and scientific literature, to accelerate scientific discovery and therapeutic innovation.

Over the past decade, I have led AI-driven research across drug discovery, translational science, and biopharmaceutical R&D. My current focus includes virtual cells, biological world models, and agentic AI that predict cellular responses and guide experiments, accelerating the path from data to therapies.

🧠 Foundation Models for Biology🧬 Virtual Cells & World Models🤖 Agentic AI for Science✨ Generative Biology🔗 Multimodal Learning
✉️ Get in touch 🎓 Google Scholar 💼 LinkedIn

Overview of my research, from multimodal representation learning toward virtual cells

👤 Short Bio

I received a master's in Bioinformatics from the Max Planck Institute for Informatics and Saarland University in Germany, and did my PhD in Computer Science with the International Max Planck Research School for Computer Science at the Max Planck Institute for Informatics, where I was fortunate to be advised by Prof. Marcel Schulz. I was also a visiting student in Prof. Ernest Fraenkel's lab at MIT.

From April 2019 to February 2021, I was a postdoctoral research associate at the Broad Institute of MIT and Harvard and MIT CSAIL under the supervision of Prof. Manolis Kellis. Today, I am a Principal AI Scientist at J&J.

⚽ In my spare time, I play soccer and go hiking. My favorite soccer player is Cristiano Ronaldo.

📰 News
Oct 2026PaperOur work Spatially Informed and Direction Sensitive Representation Learning for Spatial Transcriptomics was accepted at the NeurIPS 2026 AI4DD Workshop! 🎉
Oct 2026PaperOur work Disentangling Marginal Capacity from Geometry in Optimal-Transport Label Transfer was accepted at NeurIPS 2026! 🎉
Oct 2026PaperOur work VCLMU: Mechanism-Centric Virtual Cell World Modeling for Perturbation Response was accepted at NeurIPS 2026! 🎉
Oct 2026TalkInvited talk on Scaling AI across Pharma R&D at AIDDD. 🎤
2026PaperOur work PaSTel, bridging histology and spatial transcriptomics through multimodal, bio-prior-guided contrastive pretraining, was published at ICML 2026! 🎉
Jun 2026TalkInvited talk on AI in Biotech: Accelerating Drug Discovery & Development at Future Biotech Expo. 🎤
Oct 2025TalkInvited talk on AI's Impact on the Future of Biotechnology at MIT Department of Engineering. 🎤
2025TalkInvited talk at Boston Biotech Summit 2025 on how AI is transforming value creation in biotech (see blog about my talk). 🎤
Jul 2025PaperPresented our paper on a multimodal foundation model for gRNA activity prediction at ICML 2025, Vancouver (full text). 🎉
2024TalkInvited talk at Boston Biotech Summit 2024 on AI applications in the biotech and pharmaceutical industry (watch my pre-summit interview). 🎤
🎓 Students & Interns
Ali Khodabandeh
University of Central Florida
Multimodal foundation models for spatial omics and scRNA-seq
Yuwei Miao
University of Texas
Foundation models for virtual cells
Junchao Zhu
Vanderbilt University
Multimodal foundation model for spatial omics
Sophie Guo
MIT
Generative modeling of perturbational single-cell data
🏆 Awards