Senior Software Engineer, Agent Platform (AI for the Planet)

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 $146,880 - $220,320, and in addition we have generous bonus plans to provide a competitive compensation package. Who We Are: We are a small engineering team at the Allen Institute for AI, part of AI for the Planet. Our organization works on maritime conservation, food security, disaster resilience, and climate solutions with some of the most impactful organizations on the planet. This team focuses on how we build, ship, and run agentic tools across multiple programs. We work through infrastructure, security, evaluation, user-interfaces, and monitoring for workflows that are highly non-deterministic, and we form product opinions about a field that is still taking shape. Today, our team supports a few applications: EarthRanger helps conservationists make informed operational decisions. By integrating real-time data from hardware sensors, field personnel, satellite imagery, and other sources, we help prevent the poaching of endangered species, fight deforestation, promote human-wildlife coexistence, and advocate for legal protection of wildlife. From iconic national parks throughout Africa to species-specific conservation programs in the arctic, today over 1000 conservation missions in 100 countries are using EarthRanger to protect animals and their habitats. Learn more at https://earthranger.com . Skylight uses AI to detect illegal, unreported, and unregulated fishing in real time. Governments, enforcement agencies, and conservation organizations in 95+ countries use it to protect their waters. Our advanced AI-powered platform delivers real-time vessel detections and actionable insights that empower enforcement agencies globally to protect marine ecosystems. Read more at https://allenai.org/skylight . OlmoEarth is an open, end-to-end platform built around our family of foundation models for Earth observation. The OlmoEarth Platform enables our users to create custom fine-tuned models to detect and classify novel geospatial features. The platform handles the full loop: imagery acquisition from Sentinel-1, Sentinel-2, and Landsat; annotation; distributed training and inference; and a viewer so the outputs are usable by people who aren't ML experts. Partners today include NASA JPL (wildfire risk), IFPRI (crop mapping in Kenya), Global Mangrove Watch, and the Amazon Conservation Alliance. Read more at https://allenai.org/olmoearth . If you're the kind of engineer who gets energized by building technology that helps protect oceans, forests, and the climate, who wants to move fast, work across disciplines, and see your code have real-world impact, this is for you. What We Believe: The mission is the point . We're building AI for the planet: environmental conservation, food security, climate. If it's important to you to work on problems with a positive impact on our planet and the world, you're in the right place. The engineer closest to the user makes the best decisions. We put weight on talking to users, sitting with partnerships, and working side by side with researchers. You can't ship the right thing if you don't understand who you're shipping it for. This engineering team travels regularly to meet with users. Iterate small . Our users are tackling huge problems: illegal fishing, food security, climate resilience. They need tools that genuinely help. We believe the fastest way to build those tools is to design and build alongside them as partners: ship something functional, learn from how they use it, and iterate from there. Keeping users in the loop is how we build a better system, faster. We ship high-quality code quickly, and we learn fast from mistakes . We hold a high bar for what we put into production, but we also move with urgency. When something breaks, we focus on understanding the system, not blaming individuals. Failures are signals that help us strengthen the layers that protect our users. In-person matters . A lot of the best work on this team happens in unscheduled hallway conversations between engineering, research, and partnerships. We're in the office most days because that's where the team is at its best. We hire for curiosity. The technologies we use will change over the years, and the engineers who do well here are the ones who enjoy learning new things, not the ones who've memorized a particular toolkit. Ideas get better when they're challenged. We make decisions by talking them through - asking questions, pushing back when something doesn't quite add up, and being open to changing our minds. Everyone here is still learning, and we like it that way. Your Next Challenge: We are building a platform for conservation groups, government agencies, and research organizations that want to run AI agents against their own systems. They know their domain and they have the data. We give them the tools to build an agent, see what it is doing, and tell whether it is any good. You would work across the whole platform, from core services to the tools, UI and API interfaces, docs, and evals authors use, and help decide what those should become as the field changes. You would also build and run agents of your own, so you know the platform as a user. Much of the job is spent with partners: meeting their developers, sometimes on site, seeing where they get stuck, and turning recurring problems into platform improvements. We expect you to form opinions about what to build, argue for them, and change them when users show you otherwise.

Build and evolve the platform's core capabilities and interfaces. Own the experience of authoring agents on the platform, from the first build all the way through production deployment. Build the UI, CLI, and SDK that authors use to deploy agents, run evals, and understand what happened