Responsibilities
- Implement algorithms for hit identification through virtual screening and other high throughput computational methods as part of a cross-functional team
- Adapt and implement cutting-edge ML architectures for co-folding to augment our extensive internal structural biology expertise and capabilities
- Design and deploy active learning frameworks that utilize experimental assay results to iteratively improve model performance and reduce the number of "Design-Make-Test-Analyze" cycles leveraging state-of-the-art de novo design, ADMET predictions, and affinity predictions
Basic qualifications
- Ph.D. preferred in Computer Science, Machine Learning, Engineering or a related field, or BS/MS + seasoned experience
- Proven experience with protein-ligand co-folding algorithms (e.g., Boltz, AlphaFold, OpenFold, etc) and the ability to integrate these structural insights into broader ML discovery pipelines.
- Advanced proficiency in Python and deep learning libraries (e.g., PyTorch, TensorFlow) is required. You must be capable of building and maintaining production-quality code and data pipelines.
- Exceptional ability to communicate the "why" behind a design to a diverse scientific audience.
Preferred qualifications
- Expert-level knowledge of deep learning frameworks, specifically for affinity prediction, ADMET modeling, and the application of LLMs in a biological or chemical context
- Expertise fine-tuning existing models with internally generated structural biology and biology data
- Experience deploying ML/AI algorithms for use by a cross-functional scientific audience
- 1-4+ years of experience in a biotech or pharma setting
- A demonstrated track record of innovation in the ML/AI space, including developing and validating new architectures or novel applications of existing models to solve complex drug discovery problems including tools for hit identification (virtual screening, HTS)
- Expert level use of protein-ligand co-folding algorithms to small molecule drug discovery ML/AI tools (AlphaFold, Boltz, OpenFold)
- Experience writing production-level code for ML tasks:
- Knowledge of key scientific packages (RDKit, scikit-learn, numpy, pandas, pytorch, deepchem, polars, PyG/DGL):
- Write robust, testable, and version-controlled code that adheres to CI/CD and data governance best practices.
- Value clarity, documentation, and structured thinking, especially when working with complex data
- Knowledge of containerization technologies (Docker, Kubernetes) and cloud deployment at scale
Tags & focus areas
Used for matching and alerts on DevFound Fulltime Machine Learning Ai