Software Engineer

Tamarind builds the inference platform for molecular AI models like AlphaFold, helping biotech and pharma scientists computationally design drugs. AI is opening up new possibilities for drug discovery, with increasingly powerful models released from research labs every week. But getting a model to run is only the beginning. Making these models useful means deploying them reliably at scale, integrating them with existing data and workflows, and meeting demanding requirements for security and governance. Pharma companies are built to discover medicines, not to build software. Tamarind builds that infrastructure, enabling research teams to put the latest models to work across their organizations without having to build the underlying software stack themselves. Today, Tamarind gives tens of thousands of scientists access to more than 300 molecular design tools through our APIs, MCP, pipelines, web platform, and AI agents. We work with 10 of the top 20 pharma companies, as well as hundreds of biotechs and academic institutions. We hit 8 figures in revenue recently and raised a $13.6M Series A. We’re still a small team, and are hiring to keep up with customer demand. We’re looking for exceptional engineers who want to solve difficult technical problems, own important systems, and help shape how drug discovery happens in the AI age. THE ROLE Our work spans distributed systems, cloud infrastructure, developer tooling, and product engineering. We deal with scaling inference across tens of thousands of GPUs, building long-horizon asynchronous agents for protein design, and maintaining secure cloud environments for our customers’ most confidential research data. You’ll own projects end to end: understand the problem and scope out requirements, make the architectural decisions, write and ship the code, and make sure the system works reliably in production. You’ll work directly with Tamarind’s founders, fellow engineers, computational biologists, and customers. We don’t have a separate team for every layer of the stack, and the most interesting problems rarely fit neatly into one category. You might spend the morning designing a distributed job scheduler and the afternoon optimizing GPU inference or reviewing agent traces. We expect engineers to find the highest-leverage problems and follow them through to a working solution. WHAT YOU MIGHT WORK ON - Scale molecular AI inference. Build orchestration systems that schedule heterogeneous CPU and GPU workloads across Kubernetes, HPC clusters, and customer-controlled cloud environments, improving scheduling, throughput, reliability, and cost. - Productionize new scientific models. Turn research repositories and newly published models into secure, reliable services that scientists can use without managing the underlying infrastructure. - Build agents for molecular design. Develop the infrastructure and tools for agents that run multi-step design campaigns, execute asynchronous workflows, evaluate candidates, and make decisions based on intermediate results. - Make complex workflows composable. Design APIs and primitives that let scientists and AI agents combine hundreds of molecular modeling tools into reproducible pipelines. - Own problems across the stack. Trace failures through APIs, queues, containers, GPU kernels, and scientific outputs until you find the underlying cause. WHAT WE’RE LOOKING FOR We care more about demonstrated ability, judgment, and learning speed than pedigree or years of experience. You may be a strong fit if you: - Have built and shipped technically ambitious software, whether at work, through research, or on your own. - Take responsibility for outcomes rather than tasks. You care deeply whether the whole system delivers an excellent experience for the user, not just whether your component works in isolation. - Can reason about large complex systems and design robust architectures, while also diving deep into particular implementation details. - Thrive in ambiguity. You can take an open-ended problem, figure out what needs to happen, and make progress without a detailed playbook. - Learn unfamiliar tools and domains quickly. You don't need a background in computational biology. - Enjoy working directly with users and translating loosely defined problems into concrete technical solutions. - Care about reliability and security without losing the ability to move quickly. - Use modern AI coding tools aggressively and thoughtfully, while maintaining a high bar for understanding, review, and correctness. - Are based in the San Francisco Bay Area or willing to relocate. TECHNICAL EXPERIENCE THAT’S USEFUL You don’t need to match every item. We’re especially interested in engineers with depth in one or more areas and the range to learn the others. - Backend and distributed systems: Python, APIs, queues, workflow engines, concurrency, and failure handling. - Cloud infrastructure: AWS, Kubernetes, Docker, infrastructure as code, autoscaling, observability, and production operations. - ML systems: PyTorch, CUDA, model serving, GPU performance, and reproducible inference workloads. - Product craftsmanship: Design intuitive APIs and interfaces with meticulous attention to detail, usability, and the overall customer experience. WHAT TO EXPECT Tamarind is an early-stage company. Priorities can change as we learn from customers, and important problems don't always fit neatly within a job description. We expect engineers to move quickly, communicate clearly, and step up when a production issue or customer commitment requires it. In return, you’ll have substantial ownership, a short path from idea to production, and direct exposure to scientists doing lifesaving research. You'll work on difficult engineering problems with a small, nimble team, and your decisions will materially shape the product and infrastructure. We care about sustained output, not performative hours. We expect people to plan ahead, communicate early when issues arise, and address the root causes of recurring problems rather than repeatedly reacting to them. If you want to build excellent software with a sharp, nimble team — and have substantial responsibility for systems that scientists depend on — we’d love to hear from you.