Output Biosciences is building a biological reasoning model that understands biology at the scale and complexity life actually operates. Their model independently learned the principles of molecular interactions, enabling drug treatments previously impossible to reach. They are generating therapies beyond traditional approaches and solving the hardest problems in AI and biology. The company is currently in stealth mode, 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. The role described involves advancing the core architecture and training of Output's foundation model, which learns biological reasoning from data. Responsibilities include designing architectures, developing training objectives, running pretraining at scale, and evaluating the model's learning. You will push forward the architecture and training objectives of our foundation model, designing approaches that are purpose-built for biological reasoning You will develop methods for the model to learn across multiple biological data modalities simultaneously, building unified representations of molecular biology You will extend the model's reasoning capabilities across biological phenomena, pushing what it can predict and understand about binding, molecular properties, and biological function You will own pretraining end-to-end: experiment design, distributed training on multi-GPU clusters, hyperparameter optimization, and iteration You will design evaluation frameworks that measure whether the model has learned real biological reasoning, not just statistical patterns in training data About You You have a PhD in computer science, machine learning, physics, mathematics, or a related field with 2+ years of post-doctoral or industry research experience, or a Bachelor's or Master's degree with 5+ years of hands-on research and engineering experience in representation learning and model pretraining You have a strong publication record at top-tier venues (e.g., NeurIPS, ICML, ICLR) with contributions to pretraining methods, self-supervised learning, representation learning, or foundation models You have hands-on experience pretraining large models on diverse heterogeneous data including designing training objectives and scaling training infrastructure You are proficient in Python and PyTorch and have experience training models on distributed multi-GPU infrastructure You have demonstrated ability to own full research-to-training pipeline: you do not just design methods you train and ship models You write production-quality code that is well-tested maintainable comfortable working shared codebases version control code review You are rigorous experimentalist who designs evaluations carefully tracks experiments systematically draws conclusions from data rather than intuition Bonus Points:

• Background chemistry biology computational biology biophysics related natural science
• Experience pretraining models molecular biological data
• Experience multimodal learning learning heterogeneous data sources
• Contributed open-source machine learning projects

Our Values: ❤️ Heart: culture ownership passionate individuals pride contributions. Excellence: unwavering commitment highest standards. Practicality: results-oriented thinking tangible impact lives patients broader community. Honesty: open transparent issue addressing. Fun: creating fun engaging rewarding workplace. What we offer:

• Encouragement new ideas creativity contrarian thinking.
• Healthy feedback-focused environment leadership support.
• Ownership day-to-day management focused milestone achievement.
• Competitive salary equity well-funded startup.
• Excellent medical dental vision coverage.