Software Engineer, Core Science

OpenAI•San Francisco, CA

8d•$293,000 - $385,000•Hybrid

About The Position

The Core Sciences team is a new vertical within Codex focused on making AI an exceptional collaborator for scientific discovery. The team is building AI-native tooling and infrastructure for scientists across paper writing, literature search, data analysis, simulations, and computational experimentation. Our mission is to help researchers move faster — accelerate breakthroughs in math, physics, biology, chemistry, bioengineering, and adjacent scientific domains through powerful AI-assisted workflows. As a Software Engineer, Core Sciences, you will help design and build the platforms, products, and infrastructure powering AI-native scientific research workflows inside Codex. We’re looking for people who are excited about building ambitious 0→1 systems at the intersection of AI, engineering, and scientific discovery. This role spans both backend and full-stack engineering, with opportunities to work closely with researchers, applied scientists, and product teams to define entirely new experiences for scientific exploration and experimentation. This role owns systems end-to-end: from architecture and implementation to evaluation, launch, and production operations, with a strong bias for both quality and velocity. You will build across backend services, data and orchestration pipelines, model-powered workflows, and user-facing experiences, making thoughtful trade-offs as scientific use cases and AI capabilities rapidly evolve. Success in this role means turning ambitious, ambiguous ideas into reliable products that researchers can use to analyze data, run simulations, explore evidence, and accelerate scientific discovery. This role is based in San Francisco, CA. We use a hybrid work model of 3 days in the office per week and offer relocation assistance to new employees.

Requirements

Have strong backend or full-stack engineering fundamentals and experience taking complex product ideas from prototype to reliable production systems, with attention to performance, usability, and maintainability.
Are comfortable building in highly ambiguous 0 → 1 environments, where customer needs, product direction, technical architecture, and model capabilities are evolving quickly.
Are excited about applying AI systems to consequential real-world scientific problems, and care about building tools that improve how researchers reason, analyze evidence, collaborate, and make discoveries.
Have experience building scalable distributed systems, developer platforms, data-intensive applications, retrieval or knowledge systems, workflow tools, or model-powered product experiences.
Care deeply about product quality, iteration speed, and user impact, and can use feedback from researchers and domain experts to rapidly improve systems while maintaining scientific and technical trustworthiness.
Are curious about scientific workflows and domains such as computational biology, biology, chemistry, biomedical research, scientific literature, or research tooling, and are motivated to learn from experts in these areas.
Communicate clearly across engineering, research, product, and domain-specialist audiences, and enjoy turning complex, underspecified problems into opinionated systems that researchers can depend on.

Responsibilities

Shape the future of AI-powered scientific research by building backend systems and full-stack product experiences across Prism and Codex
Develop AI-native workflows for paper writing, literature review, paper understanding, research synthesis, and scientific knowledge exploration, designing systems that can work across complex technical content, heterogeneous sources, and evolving research questions.
Partner closely with researchers and domain experts across biology and adjacent scientific disciplines to understand real research workflows, identify high-leverage opportunities for AI, and translate ambiguous scientific needs into useful, trustworthy product capabilities.
Build AI-powered tools for scientific data analysis and simulation, helping researchers explore complex datasets, design and run computational experiments, interpret results, and iterate on models or hypotheses more efficiently.
Help establish the product, platform, and engineering foundations for fast-moving 0 → 1 efforts across the Core Science organization, balancing rapid experimentation with the rigor, reliability, and quality required for tools used in research.