Flagship Pioneering is seeking a consulting ML Enzyme Designer to build and optimize computational pipelines that prioritize protein sequences for experimental screening in enzyme discovery and engineering.
Requirements
- Multiple sequence alignment, phylogenetics, and ancestral sequence reconstruction (ASR).
- Structural modeling and function prediction of enzymes.
- Experience with supervised, unsupervised, and generative models.
- Familiarity with active learning approaches for iterative design-test-learn cycles.
- Ability to build scalable, reproducible workflows (e.g., with Snakemake, Nextflow, or cloud-based tools).
- Proficiency in Python/R with ML libraries (PyTorch, TensorFlow, scikit-learn).
- Experience with phylogenetics and molecular evolution tools (HyPhy, PAML, FastML).
Responsibilities
- Design and implement computational pipelines for ranking and prioritizing enzyme sequences for screening.
- Incorporate sequence, structural, and functional features (e.g., conservation, diversity, active site motifs, predicted stability) into multi-criteria scoring systems.
- Apply and adapt ML/AI methods (Bayesian optimization, reinforcement learning, active learning) to iteratively refine prioritization based on experimental feedback.
- Utilize tools like PAML, HyPhy, and FastML for detecting positive selection, ancestral sequence reconstruction, and evolutionary modeling.
- Leverage knowledge of enzyme mechanism, structural biology, and biocatalysis to ensure that computational outputs map to experimentally relevant hypotheses.
- Work closely with experimental biochemists, structural biologists, and protein engineers to validate computational predictions and feed results back into model refinement.
Other
- Minimum 10 hours per week.
- Excellent communication skills for interdisciplinary collaboration.
- PhD or working toward PhD in Computational Biology, Bioinformatics, Biophysics, or related field.
- Prior experience in enzyme discovery or biocatalyst engineering.
- Track record of integrating computational predictions into experimental pipelines.
