Scientists searching for a molecule do almost no market intelligence. That's why most of those molecules die — not for scientific reasons. The agents that run commercial intelligence can follow a molecule from discovery through positioning, trial design and market choice.
Before we built agents for the commercial side, we built the science
Models
B1
Ligand-based model
Screens very large compound libraries and predicts potency, safety and manufacturability.
Peer-reviewed in JCIM 2025 · preprint arXiv:2406.14572
B2
Sequence-based model
Identifies potent, selective compounds for novel targets.
In evaluation with academic partners
B3
Universal ADMET predictor
Full-profile ADME/Tox predictions that flag liabilities early.
Benchmarked against public ADMET suites
Peer-reviewed work
Published with the data and the code where the venue allows it. Every one of these predates the commercial platform, and the same evaluation discipline runs underneath it.
Case studies
Programs run end to end on our own models, from screen to nominated candidate.
VEGFR3 (FLT4)
Case study: Discovering Selective VEGFR3 (FLT4) Inhibitor for Solid Tumors
TNIK
Case study: Discovering Oral Obesity Drug targeting TNIK.
LRRK2 (G2019S)
Case study: Discovering Best-in-Class LRRK2 Therapies for Parkinson’s Disease
Academic collaborations
Awards
Recognition for applied AI in pharma and for machine learning in biology.
LLM-Based Agents for Competitive Landscape Mapping in Drug Asset Due Diligence
We present a competitor-discovery agent for drug asset due diligence that, given an indication, identifies competing drugs and extracts their attributes — a task complicated by fragmented, paywalled, alias-heavy, and fast-changing data. Since no public benchmark exists, we built one by converting five years of biotech VC diligence memos into a structured evaluation corpus, and added an LLM-judge agent to filter false positives. Our system, Bioptic Agent, achieves 83% recall, beating OpenAI Deep Research (65%) and Perplexity Labs (60%). Deployed in production, it cut analyst turnaround time for competitive analysis from 2.5 days to ~3 hours (~20x).
NeurIPS 2024 - Predict New Medicines with BELKA
Bioptic presented its gold-winning BELKA solution at NeurIPS 2024, highlighting how our AI models navigate billions of compounds to predict protein–ligand binding with precision. The work, led by Vlad Vinogradov and based on a dataset co-developed by our team, shows how AI-native tools can transform early drug discovery.
CAFA 5 Protein Function Prediction
Bioptic’s Vlad Vinogradov strikes gold at CAFA 5, the global protein function prediction challenge. His top-ranked model decoded complex protein biology from raw sequences, proving once more that Bioptic’s AI talent is at the frontier of drug discovery and protein research.