SlipstreamJobs tracks this role from the company's public career site. Apply directly on the employer's site.
Gradient Labs is building an AI customer operations platform for financial services, working with major banking and fintech companies. Founded in 2023 by alumni from Monzo, Wise, Revolut, and Google, the company deploys specialist AI agents to eliminate manual work in customer support and back-office operations.
As an AI Software Engineer, you'll own the full lifecycle of production AI agents handling live customer conversations at scale (tens of thousands per hour). This is a hands-on, build-and-ship role where you'll design and implement agentic systems powered by LLMs, including tool calling, multi-step reasoning, and integration with customer APIs and data sources.
Key responsibilities include: translating ambiguous business problems into scoped, shippable agentic workflows; building robust evaluation infrastructure with structured evals measuring accuracy, safety, and latency; developing and optimizing agent skills through dataset curation and iterative testing; contributing to shared libraries and patterns for agent development across the platform; and analyzing customer data to identify automation opportunities.
You'll work with a strong product mindset, prioritizing customer and business value over pure ML technique. The role emphasizes fast iteration, pragmatic decision-making, and clear communication. You'll have autonomy to scope your own work, push back on framing, and drive technical decisions.
The team values hands-on experience shipping products to real customers at scale, comfort with ambiguity, and the ability to break down complex problems from first principles. No specific tech stack is required—just good judgment about what's right for each problem.
Benefits include private medical and dental insurance, 25 days holiday plus bank holidays (or 32 days with opt-out), team socials and offsites, and flexible working. The London office operates on a hybrid model; this role is open to London-based or UK-remote candidates (ideally attending the office twice per week).
