Senior Research Engineer

Persons in these roles are expected to work from our offices in Seattle. On-site requirements vary based on position and team. If you have questions about on-site work arrangements for this role, please ask your recruiter. Our base salary range is $174,240 - $261,360, and in addition we have generous bonus plans to provide a competitive compensation package.  Who You Are: OlmoEarth is growing — more partners, more use cases, and a platform that is evolving quickly. We are looking for a Senior Research Engineer who can collaborate with our partners to tailor the OlmoEarth models to a wide range of specific applications across multiple domains.  Who We Are:  OlmoEarth is an open, end-to-end platform built around a family of foundation models for Earth observation. The platform enables users to create custom fine-tuned models to detect and classify novel geospatial features, handling the full loop: imagery acquisition, annotation, distributed model training and inference, and visualization. Our partners span some of the most respected institutions working on wildfire risk, crop mapping, mangrove conservation, and forest protection. OlmoEarth sits within the AI for the Planet group at the Allen Institute for AI, a small, mission-driven team working on conservation, food security, disaster resilience, and climate solutions. Learn more: https://allenai.org/olmoearth What We Believe The mission is the point. OlmoEarth exists to put powerful Earth observation tools into the hands of people working on conservation, food security, and climate. Every partner engagement this role supports is connected to that goal. If it matters to you that your day-to-day work adds up to something larger, you are in the right place. Good operations are invisible and indispensable. When coordination, documentation, and follow-through are working well, the whole team moves faster and partners have a better experience. This role is the engine behind that. Our partners are the signal. We learn what to build and how to improve by staying close to the people using the platform. The feedback, patterns, and friction you surface in this role directly shapes what the team works on next. In-person matters. A lot of the best work on this team happens in quick, unplanned conversations between engineering, research, and partnerships. We are mostly in the office because that is where this kind of collaboration happens naturally. Say what you think. We make better decisions when people share what they are actually seeing — whether that is a process that is not working, a partner need we are missing, or an idea for doing something differently. Everyone here is still learning, and we like it that way. Your Next Challenge: You will work with partners to deploy OlmoEarth for their use cases. This will require you to move fluidly across the entire OlmoEarth team, working with partners, engineers and researchers. You will make meaningful contributions to all the components of OlmoEarth’s infrastructure (from the finetuning code to model pretraining to our rslearn backend). Use case enablement Collaborate closely with partners to deploy OlmoEarth models in challenging contexts. This will prioritize contexts and partners for which we don’t have immediate solutions or there’s an opportunity to standardize a high quality approach for common use cases..  Explore novel use cases for the OlmoEarth models (e.g. post-hoc addition of new modalities, effectively leveraging embeddings in different contexts) which can unlock new use cases and partners. Partner Communications & Coordination As part of model development, maintain communications with key partners to ensure their success using the OlmoEarth platform. Communicate  internally  so that partner needs are clearly understood by the OlmoEarth machine learning research, engineering and partnership teams.  Product Improvement and Research Work closely with the engineering and partnerships  team to feed lessons you learn when deploying models into our infrastructure. This includes improvements to rslearn, OlmoEarth Studio.  Collaborate with the research team to identify and fix issues with the OlmoEarth models preventing their deployment in specific important applications.  Continually update our model adaptation approaches to improve model performance for all our partners. This includes updating our fine-tuning approaches, developing recipes for new applications and improving the UI so that modelling trade-offs can be better understood by users.  Support agent evaluations and development. What You’ll Need: Required 2+ years of experience deploying machine learning solutions. This covers the full stack of machine learning, including understanding the business case and requirements, training models, and deploying them at scale. Technical experience using machine learning tools. This includes fluency in PyTorch, experience debugging traini