Member of the Technical Staff, Biological Data

Output Biosciences•New York, NY

About The Position

Output has built a biological reasoning model that understands biology at the scale and complexity life actually operates. Our model independently learned the principles of molecular interactions, opening up drug treatments that were previously impossible. We're already generating therapies that traditional approaches cannot reach. The hardest problems in both AI and biology are being solved here, and there is room for you to own one. Output is currently in stealth, operated by a team of repeat founders and biotech veterans with multiple exits in AI x Bio, and backed by top-tier VCs including Y Combinator. You will own the data that our models learn from. This role requires a deep understanding of molecular biology - what a biological data source contains, what it implies, and what is missing. The quality and coverage of training data determines what our models can learn, and the biological insight behind how that data is constructed is the difference between a model that memorizes and one that reasons.

Requirements

You have a PhD in computational biology, biophysics, structural biology, chemistry, biochemistry, or a related biological field with 2+ years of post-doctoral or industry research experience, or equivalent depth through a combined biology and computational background
You have deep understanding of molecular interactions, protein structure, and biological data at the molecular level, grounded in first principles rather than surface familiarity
You have experience working with large-scale biological or molecular datasets, including sourcing, cleaning, integrating, and analyzing heterogeneous data
You have strong programming skills in Python and are comfortable building computational pipelines for data processing at scale
You understand what machine learning models require from training data: coverage, quality, balance, and evaluation rigor
You approach data construction as a research problem, not a pipeline task: you think carefully about what data means, what signal it carries, and what is absent

Nice To Haves

You have experience with computational biology tools such as structure prediction, molecular docking, or virtual screening
You have experience training or evaluating machine learning models, particularly on molecular or biological data
You have publications in computational biology, bioinformatics, or molecular informatics
You have a background in cheminformatics or molecular data analysis
You have experience working with protein or molecular language models

Responsibilities

You will construct training datasets that capture how proteins and molecules interact, drawing from diverse biological data sources and extending them with your understanding of molecular principles
You will develop methods to expand training data beyond what exists in public databases, using biological and chemical reasoning to create new training signal where current data is sparse or absent
You will design benchmarks grounded in real molecular phenomena, measuring whether our models have learned biologically meaningful capabilities rather than statistical shortcuts
You will develop data strategies in collaboration with model researchers, determining what the model should learn from, what biological signal to prioritize, and how to sequence learning across modalities
You will design approaches for integrating data across biological scales and modalities, building coherent training data from heterogeneous experimental and computational sources
You will design rigorous splitting and evaluation strategies that prevent leakage and ensure model capabilities generalize to real biological problems
You will stay current with biological data sources, experimental methods, and molecular databases, continuously identifying new sources of training signal