Senior ML Engineer, Federated ADMET and Toxicity Pipelines

Apheris; United States; Remote; USD 140,000 - 210,000; Full time

Benefits offered by this job

Remote-first working

Wellbeing budget

Mental health support

Learning budget

Office days at Berlin HQ

Job summary

Apheris is seeking a senior ML engineer to own training and evaluation pipelines behind our ADMET and toxicity networks—turning partner data into trained, benchmarked models that run inside partner environments where data cannot be seen.

You will work at the intersection of molecular ML and federation, collaborating with the scientific lead for toxicity to define what we model and why while ensuring reproducibility, validation, and auditable provenance across federated runs.

Qualifications

• 5+ years building ML systems in Python with software engineering discipline.
• Hands-on molecular ML or cheminformatics using RDKit, fingerprints and descriptors, or graph/transformer models.
• Experience building training/evaluation pipelines that others run, not notebooks.
• Understanding data leakage, split design, applicability domain, dataset shift, and robust benchmarking.

Responsibilities

• Own the model pipelines end to end—from landing data to released model weights.
• Build validators and data contracts with actionable errors and QC reports for partners.
• Ensure reproducible and auditable federated runs with versioned configs and data snapshots.
• Develop leakage-safe evaluation and honest performance reporting for external partners.
• Translate scientific requirements into defined, testable pipeline behavior.
• Collaborate with scientists to surface data or modeling risks early to partners and internally.

Skills

Python

MLOps discipline

Molecular ML

Collaborating with scientists

Tools

RDKit

PyTorch

Kubernetes

Job description

Apheris is seeking a senior ML engineer to own training and evaluation pipelines behind our ADMET and toxicity networks—turning partner data into trained, benchmarked models that run inside partner environments where data cannot be seen.

You will work at the intersection of molecular ML and federation, collaborating with the scientific lead for toxicity to define what we model and why while ensuring reproducibility, validation, and auditable provenance across federated runs.