AI Research Engineer

Dayhoff Labs; Cambridge (MA); On-site; USD 150,000 - 230,000; Full time

Benefits offered by this job

Visa sponsorship

Job summary

Dayhoff Labs is seeking a researcher to design and train frontier models spanning biology and chemistry. Your work will directly influence wet-lab experiments as models predict outcomes and guide experiments.

You will own end-to-end model development, collaborate with chemists and biologists, and iteratively fold new data into each cycle to improve predictions. A PhD is highly preferred and visa sponsorship is available.

Qualifications

• Proven experience training large models end-to-end with ability to diagnose and fix issues.
• Strong ML engineering fundamentals across architectures, training dynamics, data pipelines, and evaluation.
• Experience solving chemistry/biology or physically-based problems and communicating with biology/chemistry colleagues.

Responsibilities

• Train large models across enzyme-substrate prediction, neural network potentials, and inverse design of reaction networks.
• Own models end to end: architecture, data pipelines, training, debugging, evaluation.
• Collaborate with chemists/biochemists and translate between biology/chemistry and AI concepts.
• Incorporate new data iteratively to improve models.

Skills

Large model training end-to-end

ML engineering fundamentals

Chemistry/Biology problem experience

Education

PhD in a quantitative field

Tools

Graph neural networks

Neural network potentials

Python / PyTorch

Job description

About us

We're reverse-engineering the origin of life - one of the great unsolved problems in science, and one we think AI finally makes tractable.

If we succeed, the applications are vast: from catalysis and green synthesis to ab initio synthetic biology and programmable matter. Understanding and harnessing these processes could let ten billion of us thrive on this planet - and let us dream that diverse life keeps evolving and thriving beyond it.

We're a small, diverse team of AI engineers, computational scientists, and bench scientists. We hold ourselves to the rigor of a research institute, but we ship like an engineering firm. Global team, HQs in Cambridge, MA and London, UK.

The role

You'll design and train frontier models across biology and chemistry - molecular structure and dynamics, reactions, whole reaction networks. What your models predict decides what the wet labs run next; what the labs find decides your next model. We'll back an approach that might not work if the upside is large enough.

What you'll do

• Train large models across three threads: enzyme-substrate prediction, neural network potentials, and inverse design of reaction networks

• Own models end to end - architecture, data pipelines, training, debugging, evaluation

• Work directly with chemists and biochemists, and translate between the two fields fluently

• Fold new data into each iteration

Essential experience

• Demonstrated experience training large models end to end, with the depth to discuss in detail what broke and how you fixed it

• Strong ML engineering fundamentals: architectures, training dynamics, data pipelines, and evaluation

• Prior experience working directly on a chemistry, biology, or related physical-science problem, combined with the ability to communicate complex technical concepts clearly to colleagues whose first language is biology or chemistry rather than AI

Highly preferred

• PhD in a quantitative field (common on our team, but depth matters more than credentials)

• Experience in ML-for-chemistry or ML-for-biology (e.g., neural network potentials, graph neural networks, protein or reaction models)

Logistics

Compensation is highly competitive. We're also able to sponsor visas for the right candidate.