Software Engineer
Fractional AI is seeking a Staff Software Engineer for their Dubai office. This role is ideal for experienced engineering leaders or founders who want to remain individual contributors, focusing on end-to-end technical outcomes and writing production-ready code for AI products. The engineer will spend 75% of their time coding and 25% working directly with clients to understand problems and ensure solutions meet their needs.
Fractional AI is focused on putting frontier AI to work. We’re a group of veteran Silicon Valley builders who care deeply about getting complex AI systems built right, with strong conviction about what makes them succeed. We're backed by Anthropic, Blackstone, Hellman & Friedman, Goldman Sachs, General Atlantic, Leonard Green & Partners, Apollo Global Management, GIC, and Sequoia Capital. We're headquartered in SF with offices in NYC, Raleigh-Durham, and Dubai.
This role is designed for former engineering leaders (IC or EM) or founders who are comfortable owning end-to-end technical outcomes but specifically want to continue being impactful as individual contributors and spend more time in the code.
You'll work on a small, high-caliber team (2–5 engineers and a PM) building AI products for our clients. You'll set technical direction, write code, and be the person the team looks to when something is hard.
You'll spend roughly 75% of your time in code and 25% working directly with clients — often their CTOs — understanding problems, walking through tradeoffs, and making sure what we’re building meets their needs.
Most engineers take a decade to see this range of hard problems across this many domains. Here, you'll do it in your first year. To make this possible, we are religious about being the best company in the world to learn Applied AI engineering practices.
We're opinionated about what works in production, and we write about it — from why most eval setups fail by collapsing everything into a single score, to why defaulting to chat interfaces limits AI's impact.
We have a distinct approach to building AI systems: applying standard software engineering discipline to the non-deterministic world of frontier AI. We design systems around testable hypotheses, curate durable data sets, and ensure what we ship keeps working long after we've handed it over.
After shipping 30+ AI products, we have strong opinions on what makes AI products succeed and how to get them into production. We think you'll find it refreshing if you've seen how most companies approach AI.
A typical engineer here ships two to three products per year and learns from dozens more. You have outsized autonomy on each product — but you won't have a year to polish a single system. The upside is constant time on the frontier of models and tooling, and a muscle for AI product development that's hard to build anywhere else.
Most engineers look back at this as the biggest growth period of their career. But it's possible you won't like it — we recommend using the interview process to hear what our engineers think.
Posted June 8, 2026