Machine Learning Scientist

Flagship Pioneering, Inc.•Somerville, MA

1d•$132,000 - $209,000

About The Position

We seek a Machine Learning Scientist to join our ML team and lead focused efforts in applying and adapting proprietary foundation models to enable Cellarity’s predictive drug discovery platform. As a lead ML Scientist the candidate will develop and apply AI methods to identify novel interventions and targets to accelerate early drug discovery. This role involves hands-on modeling of high-dimensional biological data, leveraging and fine-tuning state-of-the-art foundation models, while enabling interpretability and biological reasoning. Ideal candidate will have demonstrated application of deep learning and computational biology to biological problems. The successful candidate will work closely with researchers in biology, chemistry, and omics technology in a collaborative environment.

Requirements

PhD in Computer Science, Computational Biology, or related field, OR Master's degree with 3+ years or Master's or Bachelor's degree with 6+ years of relevant ML research experience for drug discovery.
Strong foundation in statistics, deep learning and generative AI.
Experience with high-dimensional biological data analysis (bulk/single-cell RNA-seq, gene regulatory networks, PPI networks, multi-omics integration).
Experience with chemical / CRISPR perturbation screen data (e.g., Perturb-seq), including analysis and modeling.
Familiarity with single-cell foundation models (e.g., Geneformer, scGPT, scFoundation) and their downstream applications.
Experience applying or fine-tuning pretrained foundation models or deep generative models for downstream biological tasks.
Experience with cloud computing (AWS/GCP) and MLOps best practices.
Excellent communication skills and ability to work in interdisciplinary teams.

Nice To Haves

Experience building agentic AI systems (e.g., LLM-based agents, tool use, multi-step reasoning workflows) for scientific discovery or data analysis.
Familiarity with target identification and prioritization in drug discovery.

Responsibilities

Apply, fine-tune, and post-train foundation models (e.g., transformer-based, diffusion, VAE architectures) on single-cell RNA-seq data and other modalities to model disease cellular states
Build state-of-the-art perturbation models using multi-modal perturbation data (small molecules, CRISPR, cytokines) and phenotypic data, with emphasis on CRISPR screen data (e.g., Perturb-seq).
Develop mechanistic interpretability methods to infer gene networks and regulatory mechanisms via attention, graph-based, and/or causal representation methods, supporting downstream applications such as target identification.
Deploy and run inference on generative AI models using proprietary datasets on cloud platforms.
Establish clinically relevant benchmarking and evaluation frameworks to assess context generalization and guide model improvements.
Stay current with the latest research in foundation models, representation learning (across biology, NLP, vision, and audio), and perturbation modeling.
Collaborate with interdisciplinary scientists from biology, chemistry, and technology teams to translate research questions into cutting-edge ML solutions.
Opportunity to collaborate with and co-develop platform modules alongside other Flagship Pioneering companies.
Communicate technical concepts clearly to diverse scientific audiences.