Applied AI Engineer

ABOUT THE ROLE We’re hiring an Applied AI Engineer to work directly with Recursive’s earliest partners and users, turning frontier AI capabilities into useful applications. You’ll sit at the intersection of software engineering, research, product, and customers, building what is needed to solve real problems and helping shape how our technology is used. You’ll own technical delivery from discovery through implementation and evaluation, working closely with researchers and engineers as our automated research capabilities become more productized. Early projects will require substantial collaboration with those teams. Your work will create capacity for research while turning partner needs into working systems and reusable improvements. WHERE YOU CAN MAKE AN IMPACT - Partner outcomes: Understand the underlying problem, choose a valuable use case, and define a concrete technical plan and success criteria. - Hands-on engineering: Build prototypes, integrations, agents, and dependable applications using Recursive’s models and research tools. - Research and product feedback: Turn deployment experience into better tools, evaluations, and reusable capabilities for the wider team. WHAT YOU’LL DO - Lead technical discovery with early partners, explain relevant capabilities and limitations, and scope proofs of concept around measurable outcomes. - Design, implement, debug, and evaluate solutions across the stack, using model APIs, agent harnesses, retrieval, or tool integrations where they fit the problem. - Own delivery from problem definition and architecture through a working prototype and, where appropriate, production deployment; manage milestones, dependencies, and technical risks. - Use Recursive’s research tools to run partner projects, bring researchers and engineers in where their expertise is needed, and give precise feedback on failures and missing capabilities. - Build evaluations and instrumentation that test the actual use case; make deliberate tradeoffs among quality, reliability, latency, cost, and delivery time. - Synthesize technical results with go-to-market colleagues, communicate outcomes and next steps to partners, and bring reusable code and patterns back into our shared tools. WHAT WE’RE LOOKING FOR - Strong software engineering fundamentals and experience personally building, shipping, and maintaining production systems. - Strong programming and debugging skills in Python, TypeScript, or a comparable stack, with the ability to learn unfamiliar code and verify that changes work. - Practical understanding of modern AI systems and how model behavior affects application design, evaluation, and reliability. - Product judgment: you can discover the real problem, assess feasibility, prioritize the work, and make pragmatic decisions when requirements are incomplete. - Clear technical communication and strong interpersonal skills, including the ability to build trust with customers and align technical and nontechnical stakeholders. - High agency and end-to-end ownership. You take initiative, work through obstacles, and stay close to both the code and the user as priorities change. ESPECIALLY RELEVANT EXPERIENCE - Experience as a technical founder, early startup engineer, product engineer, or customer-facing engineer with substantial implementation ownership. - Hands-on work with LLM applications, agents, retrieval, evaluation frameworks, model deployment, or integrations with existing systems. - Partner projects that moved from an ambiguous problem to a validated application, with clear examples of what you built and learned. - Working closely with research teams to apply emerging capabilities. Experience training or changing model weights is useful but is not required. WHAT SUCCESS LOOKS LIKE Partners receive working solutions with clear evidence of their value and limitations. You keep delivery moving, use researchers’ time effectively, and communicate results clearly. The work leaves behind reusable tools, stronger evaluations, and actionable feedback that helps Recursive’s research and products improve.