Systematic Oversimplification

Swagatam Mukhopadhyay (Swɑːg), PhD

Cross-Disciplinary Serial Entrepreneur: Applied Mathematics, Theoretical Physics, modern AI, Quantitative & Molecular Biology, Genomics, Nucleic Acid Drug Development & Personalized medicines

I love teaching AI the world of molecules and Quantum Physics. The Physical AI industry is focused on Newtonian Physics. Creating better matter will need AI to learn the rules of the quantum world.

A polymath best describes my career. My PhD was in Condensed Matter Physics. I have since travelled across multiple domains — AI, Genomics, Neuroscience, Quantitative & Molecular Biology, Pharmacology & Drug Development— over a career that spans nearly 3 decades of academic and industry research & leadership in high-performance cross-disciplinary teams.

In 2019, I co-founded Creyon Bio and co-led it (from Angel to Series A1) as the Chief Scientific Officer & Board Member. We raised significant capital from Deep Tech investors and scaled the company from zero to over seventy employees.

My mission with founding Creyon was to make gene-centric medicine immediately accessible and affordable globally. The crazy What if?

What if we could build a drug printer that would sit next to a sequencer to print clinic-ready medicine immediately, targeting the specific genetic cause (mutation, pathway) of a diseases? What if we could predict human pharmacology of nucleic acid drugs with such high accuracy that Phase I/II becomes just a validation step, and not a trial-and-error human experiment?

I am big believer in the informational drug; oligonucleotides (ASOs, siRNAs, aptamers, mRNA; all polymeric molecules by the way) as the ideal gene-expression modulators. Over the years, I have also worked on gene-editing, mRNA modalities and peptides.

At Creyon — a nucleic acid engineering company — our team went from bed-to-bedside in creating and dosing a novel N-of-1 personalized medicine in a record time of under 13 months. We saved the life of child.

I have built wet labs and deeply care about the data generation process. At Creyon, I hired and led the wet lab team, especially in San Diego & India and built an extensive CRO network spanning chemistry, biology and supply—novel nucleic acid chemistry R&D, oligo synthesis, animal pharmacology, GMP/ GLP etc. I also led the dry labs in Theory, AI/ML & Software Engineering. My team created AI models that mapped quantum chemical features to all the optimal clinically-relevant pharmacology measures of nucleic acid medicines. I also led development of NGS methods for high-throughput quantification of the dose-response of oligonucleotides. I led the creation of novel high-throughput assays that were maximally informative for cytotox, neurotox and immunotox of oligonucleotides and nucleic acid medicines.

Recent work

Over the last 2 year, I have been researching protein Language Models and RNA language models. With an amazing collaborator, Mausum, I have been investigating outlier features in LLMs.

I have been working on a stealth-mode startup with the amazing Todd Martinez, David Pekker and Stefano Ermon! We built a Polymer Foundation Model (Diffusion Transformer) for engineering all polymeric molecules and their formulations. It’s the first such generative model of its kind to predict the thermal conformations as a function of sequence and physical temperature. We extended the modeling paradigm to multi-molecule systems to solve for formulation and functionalization, with any physical state variables like concentration, solvent properties etc. as dependent variable for generation of physical states.

Also, I have been happily building glass-box reasoning agentic systems to create an AI medical assistant, with an initial focus on an AI Physical Therapy (PT) Assistant. I created a novel Hierarchical Hyper-Graph (HHG) for organizing knowledge, tracking evidence and guaranteeing agentic verifiability & safety governance.

I have strong experience in working as a consultant on business strategies, Nucleic Acid Medicine R&D, N-of-1 and N-of-few drug development and hiring high-performance ML/AI teams.

I have co-lead major business deals.

I love working on foundational, systems-level and complex quantitative problems. On the technical side of modern AI my research interests span Diffusion Models, Latent Space Methods, Attention mechanism, Transformers & Protein- RNA- Foundation Models.

I love interdisciplinary science and communicating science—at CSHL I co-led a Demystifying Science series of lectures for non-scientific staff.

How to engage

If you are interested in my expertise in Biotech + AI + Entrepreneurship reach out for consulting engagements.

If you are Physicist working in the field of AI, I would love to get to know you. If you are a founder, I have a soft spot for you and very happy to help!

If you are interested in my expertise in nucleic acid drug development (Anti-Sense Oligonucleotides, Splice-Switching Oligonucleotides, Allele-Selective Oligonucleotides, siRNAs, nucleic acid chemistry, genomics, pharmacology assays) reach out for consulting engagements.

I am selective with my engagements in N-of-1 / N-of-few cases. The field is mired in poor trial-and-error designs and severe lack of public awareness in scientifically-grounded efficacy & safety design principles / assay endpoints to create the optimal molecule! Happy to consult in accelerating the development of such personalized medicines.

Selected Press

• TNPO2 story, AI-Guided ASO Development for an Ultra-Rare Diseases
• Long Island Boy, 2, Gets ‘Life-Changing Dose Of His Own Medicine’
• Bio Builders Interview, June 2023, Builders Episode 16
• C&EN: March 2022, Creyon debuts to disrupt oligonucleotide research
• Fierce Biotech, March 2022, Creyon joins wave of AI biotechs with $40M to create ’on demand’ RNA-based medicines
• Fierce Biotech, April 2025, Eli Lilly pens Creyon Bio AI oligonucleotide pact with $1B in biobucks on the table
• Axios, April 2025, Creyon Bio inks $1B licensing partnership as it eyes more funding

Academic & Industry Research Background

• Worked on theoretical models of Vulcanization Transition, Random Solids and Polymer Systems for my PhD at University of Illinois at Urbana-Champaign — PhD Advisor was Paul Goldbart. I was enormously influenced by another polymath, Nigel Goldenfeld and a Nobel Laureate Tony Leggett who taught us Renormalization Group Theory and Quantum Field Theory, respectively
• Worked on Glassy Physics, Jamming Transition, Complexity and Disorder for my post-doc at University of California Santa Barbara— PI was Jim Langer
• Switched to understanding complexity and systems-level principles in Molecular and Quantitative Biology—working very closely with experimentalists at Rutgers University with another polymath physicist, Anirvan Sengupta
• Spent a few years at Cold Spring Harbor Lab—worked with Greg Hannon on RNAi, with Partha Mitra on large-scale Mouse-Brain Connectome project, got terribly interested and ended up working on Autism Genomics (Simons Simplex Collection) with Mike Wigler
• Worked at a large scale genotype-phenotype mapping company, Human Longevity, during the early days when it was led by Craig Venter
• Build and led the creation of the first set of ML/AI models to predict pharmacology (efficacy, tolerability, IC50) of ASOs from in vitro and in vivo (rodent) screening data at Ionis Pharmaceuticals across all critical pharmacology endpoints
• Cofounded and lead (as Chief Scientific & Innovation Officer, Board Member) Creyon Bio. I fundraised from top-tier Deep Tech VCs and was instrumental in forging 2 major business deals.
• Honed my professional management skills by completing the Executive Program for Senior Life Sciences Leaders (Mini-MBA) at Harvard Medical School—an eight months long program focused on educating senior industry professionals evidence-based management principles.

Patents

• F. C. Kamme, H.-H. Bui, S. M. Freier, and S. MUKHOPADHYAY, Compounds and methods for reducing Ifnar1 expression, US Patent App. 18/570,839, Jun. 2025
• S. MUKHOPADHYAY, Machine-learned pharmacology optimization, US Patent 12,374,431, Jul. 2025.
• S. MUKHOPADHYAY and C. E. Hart, Oligonucleotide-based machine learning, US Patent 12,057,197, Aug. 2024.