Software Engineer

About the company

At Swift Solar, our mission is to unlock the full potential of solar energy. We believe solar can provide affordable, reliable, resilient, clean energy to virtually anyone, anywhere. But while today’s solar technology is good, it’s not good enough.

Swift Solar’s breakthrough perovskite-silicon tandem technology will drive solar power to commanding new heights. Our tandem cells stack a perovskite layer on top of a heterojunction (HJT) silicon cell, delivering dramatically higher efficiency than traditional silicon alone. In 2026, we acquired Meyer Burger’s HJT intellectual property and manufacturing assets, giving us the world’s most extensive HJT patent portfolio and a clear path to gigawatt-scale domestic production.

This vertically integrated approach means we can design, manufacture, and scale next-generation solar solutions entirely in the United States—from silicon foundations to perovskite innovations. Our technology enables applications ranging from space solar to utility-scale power generation, creating American-made solutions for a world facing growing energy challenges.

Swift Solar is on a journey to bring revolutionary clean-energy technology to market. We’re in it for the long haul, with support from world-class investors. Join our extraordinary team building the cutting-edge solar technology that will power the future.

A day in the life

You'll build AI products that assist scientists and accelerate our technical progress towards next-generation solar solutions. Every experiment we run costs tool time, materials, and scientists' attention, so choosing what to run next is one of the biggest levers on how fast our cells improve. You'll build the AI that helps our scientists make those choices. Our systems read scientific literature and query our lab and manufacturing data, then propose experiment plans that an automated optimizer uses to execute them.  You will lead development of the planning and reasoning layer, which takes a research objective as input and uses a multi-agent system to guide our scientists and reach goals faster through the best selected experiment campaigns. Additionally, we need to measure what works and how frontier models and our own agents perform on problems our team has already worked through, so we can improve the system over time. 

This is also a place to experiment. When a promising new paper, model, or agent design comes out in AI for science, you can try it here on real problems instead of benchmarks someone else chose. You'll work with scientists who have spent years on perovskites, who know which questions matter and can tell you quickly whether an agent's answer makes sense. Because we develop the process and make the devices ourselves, the data, tools, and people who understand them live in the same company, so an idea can go from your screen to a measured result without waiting on anyone outside.