AI Engineer

Retire the ERP. Light exists to replace factory-era ERPs with software that feels alive. Our Smart Financial Platform gives modern, global companies superpowers—automated accounting, real-time reporting, and financial flows that move at the speed of the business. We build with our customers, ship fast, and obsess over craft. In a short time, Light has gone from idea to the operating core for leading companies like Lovable, Legora, and Keyshot. People don’t just use Light—they enjoy it. We’re an early team defining a new software category. Think engineers who love debits and credits, designers who care about reconciliation states, and operators who treat finance as a product. If you’re excited to modernize how the world runs money—one workflow at a time—you’re in the right place. Backed by world-class investors and advised by industry titans, we’re building category-defining products with the freedom to ship ambitiously and own outcomes. Come help us make Light the global default for next-gen finance. This AI Engineer role

We're looking for builders who can take a customer problem from idea to production, wherever in the stack it lives. At Light, AI writes code and the bottleneck has shifted from writing to thinking, so we want engineers who work end-to-end and ship things that matter. We are primarily hiring at mid-level for this role. Junior candidates who are exceptional and genuinely AI-native are welcome. Senior candidates will be considered by exception; we have strong technical leadership and the bar is high. Your qualifications:

2-6 years of experience ideally, plus... - You build things. Not prototypes, not demos. Working software in production, doing something real for real users. This is the non-negotiable. - You've changed the way you work because of AI tooling, not just added it to your workflow. You have an instinct for where it helps, where it misleads, and how to use it without losing ownership of the outcome. - You have hands-on experience building products or features on top of large language models. Agents are a bonus, but working knowledge of how LLMs behave in production is a must. - You don't need to be a specialist in everything, but you can't have a lane you refuse to leave. Our stack is Kotlin and React. If you're strong in one and curious about the rest, that works. If you need someone else to handle the other side of the feature, it doesn't. - You think about the person on the other end. You notice when something feels clunky, even if it technically works. You have opinions about what good software feels like and you act on them. - Small batches. Learn. Go again. You're not precious about your work. You'd rather put something real in front of a customer than spend another week perfecting it in isolation.