What Deep MedChem does

Deep MedChem is a Prague-based AI drug discovery company founded in December 2023 as a spin-off from The MAMA AI and i&i Prague, a startup investor and incubator connected with IOCB Prague (Institute of Organic Chemistry and Biochemistry). The company develops the CHEESE Platform — a family of AI-powered products for early-phase drug discovery comprising CHEESE Search (billion-scale 3D shape, electrostatic, and activity similarity search), CHEESE Modeller (ADMET prediction and custom QSAR training in 2 minutes), CHEESE Explorer (interactive visualization across 11+ public databases), and CHEESE Electrostatics (DFT-validated electrostatic similarity at 95-98% correlation). Its core technology combines proprietary Geometric Transformers (the en-transformer package) with learned molecular embeddings to enable sub-second similarity search across 40+ trillion molecules, with CHEESE Electrostatics processing single molecules in 3 seconds versus 3 hours using traditional DFT methods.

The platform integrates eight or more commercial and public molecular databases (ZINC15, Enamine-REAL, eMolecules, Chemspace, Molport, Mcule, XtalPi, Aurigene) and exposes multi-modal search across 2D fingerprints, 3D shape, electrostatic, and activity-based similarity. SynthonGPT2, launched in January 2026, extends search into ultra-large combinatorial chemical spaces. Revenue is generated through SaaS subscriptions, programmatic API access fees, on-premise enterprise deployments, custom model training engagements, and professional services. Go-to-market combines product-led growth (a 14-day free trial for CHEESE Search) with enterprise field sales for CHEESE Modeller, Explorer, and Electrostatics, supported by an active conference presence (ACS Spring 2025, BIO'24, BioML, BioVaria, ChemSpace webinars). The named customer base includes IOCB Prague (academic research) and an unnamed San Francisco Bay Area biotech; the leadership team comprises Jan Macek (CEO), Miroslav Lžičař (CTO), and Jiří Růžička (CBO).

Name

Deep MedChem

Legal name

Deep MedChem a.s.

Website

https://deepmedchem.com

Company type

Private

Founded year

2023

Operating status

Operating

Headcount range

1–10 employees

Short description

Deep MedChem is a Prague-based AI drug discovery company building the CHEESE Platform — Geometric-Transformer-powered molecular search, ADMET prediction, and custom QSAR tools serving pharmaceutical and biotech researchers with trillion-scale chemical space exploration.

Ownership category

akta.pro rank

Product category

AI Drug Discovery Software

NAICS

Research and Development in Biotechnology (except Nanobiotechnology) (541714), Computer Systems Design and Related Services (54151), Research and Development in the Physical, Engineering, and Life Sciences (54171)

SIC

Services-Commercial Physical & Biological Research (8731)

akta.pro primary industry

AI/ML Platforms for Drug Discovery & Experiment Optimization (HLAGAJAO)

akta.pro secondary industry

Cheminformatics & Molecular Modeling / CADD (HLAGAJAE)

Keywords

• Molecular similarity search
• AI drug discovery
• ADMET prediction
• Virtual screening
• Cheminformatics platform

Headquarters

HQ city

Prague

HQ country

Czechia

HQ region

Europe

Offices1 record

Markets served

GTM type

B2B

Offering type

Software

Cost components

Technology or R&D, Personnel, Infrastructure, Marketing or Sales, Operations

Revenue model

1. CHEESE Platform Subscriptions: SaaS subscription model for CHEESE Search, Modeller, Electrostatics, and Explorer tools. Free limited demo version available with premium tier offering additional functionality. Premium features include enhanced search quality, advanced filtering, and batch operations.
2. API Access Fees: Programmatic API access for registered users. Public API includes molecular search and ADMET property prediction. On-premise API provides advanced features including embeddings computation and custom database indexing.
3. On-Premise Deployment: On-premise version of CHEESE Search for enterprise customers requiring data privacy. Deployable on AWS, dedicated instances, or customer servers. Full CHEESE functionality with custom database indexing and model training on proprietary data.
4. Professional Services: Expert services providing hands-on assistance with drug discovery AI platform implementation. Collaborative approach tailoring support to specific goals and requirements.
5. Custom Model Training: Custom-trained models for proprietary combinatorial spaces. Onboarding early partners for internal virtual library search at scale.

Pricing tiers

Go-to-market motion3 records

Distribution channels4 records

Marketing channels7 records

Core offering

Deep MedChem develops and sells the CHEESE Platform, a family of AI-powered tools for early-phase drug discovery. The platform enables billion-scale molecular similarity search across 40+ trillion molecules using 3D shape, electrostatic, and activity-based metrics, ultra-fast ADMET prediction, custom QSAR model training, and electrostatic similarity search validated against DFT calculations. Offerings are delivered via cloud SaaS, REST API, and on-premise deployments for enterprise customers.

Product overview

Deep MedChem offers the CHEESE Platform — a family of AI products for drug discovery built around a unified molecular search and prediction engine. The core platform consists of four integrated tools: CHEESE Search (billion-scale molecular similarity search using 3D shape, electrostatic, and activity metrics), CHEESE Modeller (ADMET prediction and custom QSAR model training), CHEESE Explorer (public database navigation and visualization), and CHEESE Electrostatics (DFT-accurate electrostatic similarity at scale). SynthonGPT2 extends the platform to 40+ trillion virtual combinatorial molecules. All tools are accessible via CHEESE API (REST) or on-premise deployment for maximum IP protection.

Differentiator

Problem solved

Functional benefit

Brands

• CHEESE Platform: A family of AI products for drug discovery that accelerate the early phases of drug discovery, including CHEESE Search (billion-scale molecular search), CHEESE Modeller (ADMET prediction and custom QSAR training), CHEESE Explorer (public database exploration), and CHEESE Electrostatics (DFT-validated electrostatic similarity).

Products and services

• CHEESE Search Cloud-based molecular similarity search engine that screens 40T+ molecules from ZINC15, Enamine-REAL, eMolecules, Chemspace, Molport, Mcule, and other databases using 3D shape, electrostatic, and activity-based similarity metrics in sub-second time. Built for pharmaceutical and biotechnology drug discovery teams performing virtual screening.
• CHEESE Modeller Ultra-fast ADMET prediction and custom QSAR model training platform that trains predictive models on proprietary molecular data in approximately 2 minutes with state-of-the-art accuracy on ADMET benchmarks. Designed for computational and medicinal chemists needing property prediction.
• CHEESE Explorer Interactive visualization and exploration tool for navigating 11+ public databases including FDA approved drugs, patented molecules, natural products, and ZINC15, with zoomable chemical space visualization. Built for medicinal chemists and researchers exploring known compound space.
• CHEESE Electrostatics 3D electrostatic similarity search achieving 95-98% correlation with DFT calculations and processing molecules in 3 seconds that would take 3 hours with traditional DFT. Enables protein-ligand binding complementarity searches across billions of molecules for drug discovery teams.
• CHEESE SynthonGPT2 Advanced AI model that enables similarity search across 40+ trillion virtual molecules in combinatorial chemical spaces from major vendors, with built-in ADMETox filtering and vendor quotation capabilities. Designed for pharmaceutical discovery teams needing ultra-large-scale virtual library exploration.
• CHEESE API REST API providing programmatic access to molecular search, property prediction, batch operations, embedding computation, similarity matrices, visualization, and custom database indexing. Supports Python, CLI, and KNIME integrations for enterprise cheminformatics workflows.
• On-Premise CHEESE Deployment On-premise deployment of the CHEESE Search platform enabling complete data privacy with no internet transmission, deployable on AWS, customer servers, or dedicated instances with full functionality, custom database indexing, and proprietary model training for IP-sensitive enterprise customers.

Quantifiable outcome

• 147x improvement in virtual screening hit rates
• +6 more outcomes

Named customers2 records

Segments2 records

Ideal customer profiles2 records

Technology focussed Yes

API detail

Has API

Yes

API docs

API detail

Core technology

AI maturity

App detail

Integration1 record

AI capability7 records

Feature8 records

Partnerships

Ten partnerships are on record, tiered core, foundational and supporting.

• EnaminecoreTechnology or Integration · 20 May 2024Enamine-REAL database (5.5B molecules) integrated into CHEESE Search. Enables sub-second search across Enamine's ultra-large enumerated chemical space. Deep MedChem improved speed and accuracy of search functionality with this integration.
• eMoleculescoreTechnology or IntegrationeMolecules commercial library integrated into CHEESE Search database coverage. Provides access to eMolecules' ready-to-order compound collection for hit expansion workflows.
• ChemspacecoreTechnology or IntegrationChemspace library coverage in CHEESE Search. Includes joint webinar presenting molecular search capabilities in AI-learned chemical similarity space. Provides access to Chemspace's chemical inventory.
• MolportcoreTechnology or IntegrationMolport database integrated into CHEESE Search, providing access to Molport's commercial molecular inventory for virtual screening and hit expansion.
• MculecoreTechnology or IntegrationMcule-FULL and Mcule-In-Stock databases searchable through CHEESE platform. Provides access to Mcule's compound sourcing capabilities.
• IOCB PraguefoundationalStrategic or Co-development PartnerDeep MedChem is a spin-off from i&i Prague connected with IOCB (Institute of Organic Chemistry and Biochemistry). IOCB is a top-class medicinal chemistry institution providing research collaboration and talent pipeline. Jan Skácel leads a research group collaborating with Deep MedChem on AI/ML implementation for cheminformatics.
• XtalPicoreTechnology or IntegrationXtalPi database integrated into CHEESE SynthonGPT2 for ultra-large combinatorial space search. Access to XtalPi's AI-designed molecular library.
• molecule.onecoreTechnology or Integrationmolecule.one's synthetic accessibility data integrated into SynthonGPT2 for searching ultra-large combinatorial chemical spaces with built-in synthesizability scoring.
• Microsoft for StartupssupportingGTM or Marketing PartnerMicrosoft for Startups program partner. Provides startup support, cloud resources, and market access through Microsoft's startup ecosystem.
• AurigenecoreTechnology or IntegrationAurigene compound library integrated into CHEESE Search. Provides access to 40 billion molecules through Aurigene's drug discovery services and compound collection.

Scale indicators10 records

Recent moves7 records

Expansion highlights5 records

Direct peers

• Atomwise: AI-focused small-molecule discovery company using proprietary models (AtomNet) for virtual screening and hit identification across billion-scale libraries. Closest direct peer: both sell to early-phase discovery teams, emphasize speed and scale of chemical-space search, and target pharma/biotech partnerships.
• ChemAxon: Cheminformatics platform providing molecule handling, fingerprinting, search, and ADMET calculators widely embedded in pharma informatics stacks. Most comparable on the cheminformatics-infrastructure layer; Deep MedChem competes for similar medicinal-chemistry buyer budgets but differentiates on AI-learned 3D/electrostatics embeddings versus ChemAxon's more traditional rule-based toolkit.

Emerging players

• BenchSci: AI research platform for biology and drug discovery that helps scientists interpret experimental evidence and select reagents/compounds. Adjacent in serving bench scientists with AI-driven search and prioritization, though BenchSci focuses on biology/antibody data whereas Deep MedChem focuses on small-molecule similarity — both compete for the same medicinal-chemistry software budget.
• XtalPi: AI-driven drug-discovery company combining physics-based modeling with ML for hit discovery and crystal-structure prediction. Also a database partner of Deep MedChem for SynthonGPT2, which signals overlapping AI/chemistry capabilities; comparable in combining a combinatorial library platform with generative/screening tooling.

Broad incumbents

• Schrödinger: Public molecular simulation platform offering physics-based and AI-driven drug-discovery tools (LiveDesign, Glide, BioLuminate) to pharma and biotech. Broad incumbent with substantially larger R&D budget and installed enterprise base, but overlapping use cases in virtual screening and molecular property prediction where Deep MedChem's 3D/electrostatics search competes head-on.
• Dotmatics: Scientific R&D informatics platform (now under Insightful Science/Revvity) covering ELN, LIMS, registration, and cheminformatics across pharma R&D. Comparable on the drug-discovery informatics stack layer; less overlap with CHEESE's specific 3D/electrostatics search but addresses the same medicinal-chemistry user with workflow tooling.
• Recursion Pharmaceuticals: AI-driven phenomics and chemistry platform combining high-throughput biology with proprietary ML for drug discovery. Comparable in applying ML to early-phase discovery workflows (including in silico chemistry via the 2024 Exscientia merger) but operates at vastly larger scale and with significantly more capital.
• Insilico Medicine: End-to-end AI drug-discovery company with PandaOmics target ID, Chemistry42 generative/screening engine, and clinical-stage assets. Overlaps with Deep MedChem in AI-driven molecular property prediction and generative screening, but operates across a broader pipeline (target to clinic) and is far better capitalized.
• Cadence Molecular Sciences (formerly OpenEye Scientific): Molecular modeling and cheminformatics software suite (ROCS, OMEGA, FEP+) widely used in drug-discovery workflows for shape similarity, conformer generation, and 3D property calculation. Direct overlap in 3D shape similarity and molecular search — the area where CHEESE Search and CHEESE Electrostatics most directly compete.

Others

• MolPort: Compound-sourcing marketplace aggregating purchasable molecules from hundreds of suppliers — listed by Deep MedChem as both a partner and an indexed data source. Indirect competitor on the chemical-space search layer (MolPort customers use it to find compounds), but primary business is sourcing rather than AI tooling.

Market position

Strengths5 records

Weaknesses5 records

Competitive moat5 records

Key risks7 records

Key highlights7 records

Customer concentration

• Xx.com/deepmedchem

Revenue estimate

Valuation estimate

Number of profiles

Profiles6 records

Funding overview

Funding rounds1 record

Investors1 record

Funding detail is available on the Subscription and Enterprise plan.Contact sales →

M&A

Investments

M&A and investment is available on the Subscription and Enterprise plan.Contact sales →

Frequently asked questions about Deep MedChem