Senior software engineer, AI platform

About Watershed Watershed is the enterprise sustainability platform. Companies like Airbnb, Carlyle Group, FedEx, Visa, and Dr. Martens use Watershed to manage climate and ESG data, produce audit-ready metrics for voluntary and regulatory reporting including CSRD, and drive real decarbonization. We are looking for team members who love product-building, want to work hard at a mission-oriented startup, and will collaborate with us in shaping the culture of a growing team. We have offices in San Francisco, New York, Denver, London, Paris, Berlin, Sydney, Mexico City, and remote team members across the US and Europe. We hope that you'll be interested in joining us! The role Watershed is building the AI suite for companies to measure their emissions and decarbonize their business. We're looking for software engineers to help build the AI platform that powers our agents product. You'll be a technical leader laying the foundations for agentic AI at Watershed — designing the orchestration layer, controls, and tooling that let our product teams ship reliable, observable AI features on top of a wealth of operational sustainability data. In this role you will: Design and build the agent infrastructure that powers Watershed's products Develop the observability and tracing layer for agent decisions, making it possible to debug, evaluate, and improve agent behavior at scale Build evals, harnesses, and guardrails that turn agent capabilities into production-grade, dependable systems Collaborate with product and other AI engineering teams to set product and technical strategy, and define the boundaries between autonomous agent behavior, deterministic code, and human oversight Keep up with developments and state-of-the-art in AI and agent infrastructure to determine what is relevant to Watershed Work closely with Watershed product teams to contribute your expertise to build agent experiences across the product Write performant, well-crafted, tested, and maintainable code across our technical stack You might be a good fit if you have: 6+ years of experience in backend, platform, or AI/ML engineering Experience building products and infrastructure that leverage LLMs, embeddings, and other ML technologies Full lifecycle experience building, deploying, and monitoring production systems that depend on LLMs or other ML technologies Experience with model evaluation, agent observability, and making non-deterministic systems reliable Experience building and operating production Typescript systems Must be willing to work from an office 4 days per week (except for remote roles) Watershed has hub offices in San Francisco, New York, London, and Mexico City and satellite offices in Denver, Sydney, Paris, and Berlin. Where we have offices, employees are expected to be in office for 4 days per week. Certain jobs are open to being remote and will be specifically noted on the jobs page and in the job description if so. What’s the interview process like? It starts the same for every candidate: getting to know the team members through 1 to 2 conversations about Watershed, your experience, and your interests. Next steps can vary by role, but usual next steps are a skill or experience interview (e.g. a coding interview for an engineer, a portfolio review for a designer, deeper experience call for other roles) which leads to a virtual or in person interview panel. We prioritize transparency and lack of surprise throughout the process. What if I need accommodations for my interview? At Watershed, we are dedicated to ensuring an inclusive recruitment process. We provide reasonable accommodations for candidates with disabilities, long-term conditions, mental health needs, religious observances, neurodivergence, or pregnancy-related support requirements. If you need assistance during your process, please contact your recruiter.