Senior ML Engineer – ADMET & Toxicity Networks

Role overview A senior, hands-on engineering position within a federated AI platform for pharmaceutical R&D. The role focuses on building the training and evaluation pipelines that turn partner datasets into production-grade molecular property and toxicity models, working alongside scientists at major pharmaceutical companies. Code runs inside partner environments on data the team cannot directly see, so reliability, reproducibility, and scientific rigor are central.

Responsibilities - Own end-to-end model pipelines for ADMET and toxicity endpoints, from data preparation through training, evaluation, and release of benchmarked weights. - Build partner-facing validation, including schemas, data contracts, validators with actionable errors, and QC or profiling reports that work without raw data access. - Make federated runs reproducible and auditable through versioned configs, pinned data snapshots, and full provenance for every released model. - Design evaluation that holds up under scientific scrutiny, including leakage-safe splitting, held-out benchmarks, and honest performance reporting. - Harden research prototypes into tested, modular systems ready to hand off for scaling. - Partner closely with the science team to translate ambiguous scientific requirements into defined, testable pipeline behavior, and surface data or modeling risks early.

Requirements - 5+ years building ML systems in Python with strong software engineering discipline: version control, tested modular code, code review, and clean interfaces. - Hands-on experience with molecular machine learning and an understanding of how it can fail, including data leakage, split design, applicability domain, dataset shift, and over-optimistic benchmarks. - Ability to write validators and data contracts that hold up under partial visibility, where the underlying data cannot be directly inspected. - Comfortable with PyTorch or an equivalent modern ML stack. - Track record of working well with scientists, turning ambiguous requirements into defined, testable pipeline behavior. - Excellent written and spoken English.

Nice to have - Federated learning, privacy-preserving ML, or other multi-party training environments. - ML Ops or ML infrastructure experience, particularly Kubernetes-based training, evaluation, or deployment. - Production-grade model delivery in regulated, enterprise, pharmaceutical, or biotech settings. - Familiarity with public ADMET, toxicity, and bioactivity resources such as ChEMBL, Tox21, and ToxCast, including their common pitfalls. - Open-source contributions or a publication record in cheminformatics, molecular ML, or applied machine learning.

Benefits and work setup - Remote-first working model. - Wellbeing budget, mental health support, home-office and co-working stipends, and a learning budget. - Generous holiday allowance. - Industry-competitive compensation including early-stage virtual share options. - Three in-person office days per year at a European headquarters or another European location.