Young Investigator Open Language Models for Biology
Seattle-based nonprofit AI research institute developing openly released models, datasets, tools, and AI applications for science, the environment, and robotics.
Funding history
About Allen Institute for AI (Ai2)
Ai2 is a nonprofit AI research institute founded by Paul Allen in 2014. It develops openly documented language and multimodal models, training data and code, plus applied AI systems for scientific discovery, planetary intelligence, and conservation.
Skills
Candidate Availability
Required and preferred rules are kept separate and reflect the wording in the original posting.
About the Role
You will develop multimodal language models that combine single-cell gene-expression data with text. You will reproduce, evaluate, and improve relevant approaches for brain and neurodegeneration data; design controlled experiments; develop donor and brain-region representations; interpret results with scientists; and release research artifacts for publication.
Requirements
- Completed PhD or PhD expected within one year in machine learning, computational biology, computational neuroscience, physics, or a related quantitative field
- Experience training or fine-tuning transformer-based language models
- Python skills
- PyTorch skills
- Experience writing reliable, reproducible research code
- Ability to conduct independent research on open-ended problems
- Ability to communicate clearly with machine-learning researchers and biological scientists
Responsibilities
- Develop a multimodal language model combining single-cell gene-expression data with text
- Reproduce and evaluate relevant approaches, including scGPT, Geneformer, CellWhisperer, and C2S-Scale
- Adapt and improve approaches for brain and neurodegeneration data
- Design controlled experiments comparing language-grounded models with baselines
- Develop brain-region and donor-level representations incorporating pathology and disease progression
- Evaluate models on held-out donors
- Interpret results and identify predictions for experimental follow-up
- Release data, code, and model weights and prepare results for publication
Benefits
- Medical, dental, vision, and employee assistance program coverage for team members and families
- Health savings account, healthcare reimbursement arrangement, and healthcare and dependent-care flexible spending accounts
- 401k plan
- $125 per month for commuting or internet expenses
- $200 per month for fitness and wellbeing expenses
- Up to 10 sick days, 7 personal days, 20 vacation days, and 12 paid holidays per year
- Annual bonuses
- Family leave
