Head of Platform and Machine Learning, Tyler Umansky, to deliver Day 1 session “Breaking the Throughput Barrier” and join Day 2 panel “From Target to Therapy: Navigating the Drug Discovery Pipeline”
SAN DIEGO — Model Medicines, an AI-first biotechnology company developing first-in-class therapeutics against multi-indication biological choke points, today announced that it will present at the 7th ACE Drug Discovery Summit, May 6–7, 2026, in San Diego. Tyler Umansky, Head of Platform and Machine Learning, will deliver the Day 1 session on ultra-large virtual screening (ULVS). He will also join a Day 2 panel on the drug discovery pipeline.
Session Details for the 7th ACE Drug Discovery Summit
Presentation: Breaking the Throughput Barrier: Ultra-Large Virtual Screening as a Precision Amplifier for AI-Driven Drug Discovery
Speaker: Tyler Umansky, Head of Platform and Machine Learning, Model Medicines
Date: Wednesday, May 6, 2026
Panel: From Target to Therapy: Navigating the Drug Discovery Pipeline
Moderator: Rameshwar Kadam, Senior Scientist II, Johnson & Johnson Innovative Medicine
Date: Thursday, May 7, 2026
Umansky's session will frame throughput as a design choice, not a compute decision.
“Throughput should be treated as a first-order design variable, not a downstream consequence," said Tyler Umansky, Head of Platform and Machine Learning at Model Medicines. "Expanding library size and operationalizing it on accessible hardware changes the equation, improving enrichment, hit quality, and cost as we scale.”
Ultra-Large Virtual Screening (ULVS): Architecture and Training
The barrier to ultra-large virtual screening is largely economic. Every additional molecule requires another pass through a large AI model. At scale, the compute cost on premium GPU hardware quickly caps how large a screen can be. Model Medicines' GALILEO™ platform takes a different approach. The model trains on GPUs for accuracy but runs on widely available CPU infrastructure for the screen itself. This configuration delivers orders-of-magnitude higher throughput at a small fraction of the cost, enabling trillion-scale inference.
High throughput is only valuable if the model behind it surfaces novel chemistry. A common pitfall in AI drug discovery is the reliance on massive datasets to drive results. This frequently leads to overfitting, producing compounds that are iterations of known data rather than novel discoveries. To optimize for novelty, GALILEO™ trains for extrapolation. Its training set is deliberately smaller and chemically diverse, ensuring that rare scaffolds and underrepresented pharmacophores meaningfully shape the model. This emphasis on diversity over volume is what allows an ultra-large virtual screen to return unique pharmacophores rather than re-ranking known molecules. This approach also ensures the platform identifies novel drug candidates.
From the 325-Billion-Molecule Screen to MDL-4102
In 2025, Model Medicines completed a 325-billion-molecule screen, the first published AI bioactivity screen at the hundred-billion-molecule scale[1]. The screen produced MDL-4102, a highly potent and selective BRD4 inhibitor with no measurable activity against BRD2 or BRD3[2]. Selective BRD4 had been considered structurally infeasible without engaging BRD2 and BRD3, the cross-engagement that drove the dose-limiting hematologic toxicities of prior pan-BET inhibitors. ULVS scale made this result possible. Model Medicines identified a rare scaffold that maintains BRD4 potency while achieving the previously elusive selectivity. MDL-4102 is now in IND-enabling studies, with an IND submission targeted for 2027.
About ACE Drug Discovery Summit
The 7th ACE Drug Discovery Summit, hosted by ACE Expo, convenes drug discovery scientists, computational chemists, and AI/ML leaders across academia, biotech, and pharma for a two-day program on AI-driven target discovery, generative chemistry, lead optimization, and preclinical-to-regulatory integration. More information is available at https://acxpo.com/7th-ace-drug-discovery-summit-usa-2026/.
About Model Medicines
Model Medicines is an AI-first biotechnology company engineering first-in-class small molecules that target the biological linchpins underlying disease. The company’s research spans infectious disease, oncology, and inflammation, with programs designed around conserved molecular choke points that drive multiple pathologies. Model Medicines has discovered a direct-acting, non-nucleoside, broad-spectrum antiviral (MDL-001) and a potent, selective and novel BRD4 inhibitor (MDL-4102). Its work demonstrates how large-scale computation can uncover entirely new classes of drugs once thought unreachable. Model Medicines is advancing a new generation of therapeutics that redefine what is possible in modern drug discovery. Learn more at www.modelmedicines.com.
Media Contact
Patrick O’Neill
Head of Partnerships & Investor Relations
media@modelmedicines.com
[1] Google Cloud. Google Cloud to host second-annual Cancer AI Symposium in New York City [Internet]. New York: PRNewswire; 2025 Oct 30. Available from: https://www.googlecloudpresscorner.com/2025-10-30-Google-Cloud-to-Host-Second-Annual-Cancer-AI-Symposium-in-New-York-City
[2] Google Cloud. LA Tech Week - AI for Startups in Healthcare Lifesciences [Internet]. Venice (CA): Google; 2025 Oct 17. Available from: https://rsvp.withgoogle.com/events/hcls-la-tech-week
