Applied ML Specialist, Biology

• Build models with a clear use case: Train and iterate on models for assay prioritisation, experimental triage, and readout interpretation where the value is operational, not academic.
• Make outputs trustworthy: Design evaluation schemes, confidence signals, and failure checks that help scientists understand when to lean on a model and when to ignore it.
• Shape usable data inputs: Work with biology and data partners to improve labels, metadata, and lineage so models are based on something sturdier than convenience exports.
• Ship tools that people actually keep using: Turn one-off analyses into lightweight workflows or interfaces that slot into existing scientific routines instead of asking the team to change everything at once.

• Strong applied machine learning background with evidence of shipping useful models into real workflows rather than stopping at experimentation.
• Comfort working with noisy biological or healthcare-adjacent datasets where labels, metadata, and context are imperfect.
• Hands-on Python experience across modelling, evaluation, and readable implementation, using modern ML tooling where appropriate.
• Ability to explain uncertainty, edge cases, and tradeoffs to experimental stakeholders without either overselling or paralysing the work.
• Preference for specialist IC ownership over people management, with enough independence to set a high technical bar in a lean team.

• Experience in drug discovery, translational biology, assay modelling, or related applied scientific settings.
• Familiarity with PyTorch, gradient boosting, representation learning, or graph-based methods.
• Exposure to MLOps patterns such as experiment tracking, model registries, or scheduled retraining.
• Past work in distributed teams where clarity in written communication mattered as much as meetings.

Python, PyTorch, scikit-learn, SQL, Weights & Biases, Docker, GCP, GitHub Actions

Base salary

$150k-$185k, depending on applied scope and scientific context depth

Equity

Meaningful option grant plus annual cash bonus

Travel

Quarterly meetups and team travel covered in full

Timing

Fresh search tied to a broader rebuild of the internal modelling stack

1. 1.
   Talent screen: Conversation on compensation, remote preferences, and whether your background is closer to research ML or applied delivery.
2. 2.
   Manager deep dive: Technical discussion on modelling choices, evaluation, and how you decide what should or should not become productionised.
3. 3.
   Working session: Review a practical modelling brief and explain what you would build first, what you would ignore, and why.
4. 4.
   Panel final: Cross-functional interviews with biology, product-minded stakeholders, and data partners before references and offer.