The opportunity
We are seeking a Staff Machine Learning Scientist – Translational AI to provide technical leadership at the intersection of deep learning foundation models, computational biology, and molecular diagnostics. This ownership role drives the architecture and validation of genomic,…
What you'll do
Serve as the principal technical authority on the deployment of molecular,: genomic, and pathology foundation models applied to oncology and translational medicine questions
Engineer rigorous alignment and post-training workflows that ground: pre-trained foundation models in empirical clinical trial and molecular diagnostic data, eliminating speculative modeling assumptions
Formulate objective peer-review frameworks and deliver technical feedback to: elevate the modeling code, experimental standards, and scientific designs of the broader AI research group
Lead the post-training, parameter-efficient fine-tuning (PEFT), and: evaluation of deep sequence, multimodal, and representation learning models for biomarker discovery, molecular recurrence monitoring, and therapeutic response forecasting
Design robust fine-tuning, probing, and latent space representation analysis: workflows that extract interpretable, biologically grounded patterns from high-dimensional transformer architectures
Validate model outputs against multi-omic benchmarks and real-world outcomes,: ensuring model predictions deliver the exact deterministic accuracy required for patient tracking and clinical interventions
What they're looking for
- Build, train, and optimize advanced machine learning models utilizing: next-generation sequencing (NGS), ctDNA assays, digital pathology imaging, and longitudinal clinical metadata
- Design rigorous clinical investigation and evaluation frameworks that connect: model performance metrics (e.g., loss curves, precision-recall) directly to translational utility and real-world distribution shifts
- Systematically identify algorithmic failure modes, sources of dataset bias,: and covariate shift, implementing robust mitigation strategies suitable for regulated, clinical-facing pipelines
- Partner with Computational Biology, Translational Science, and Medical: Affairs teams to translate complex clinical requirements into clear, quantitative machine learning problem statements
