^{The role}

^{Most AI agents are demos. Ours answer the phone for some of the biggest brands in the world: more than 100 million calls a year, across 2,000+ live deployments. Someone is on the line trying to move a booking, report a fault or pay a bill, and the agent either works or it doesn't.}

^{We looking for a Forward Deployment Engineer who has built and shipped LLM agents in production, including tool use and evals. Forward Deployed Engineers are the people who make it work. You own a customer's agents from the first scoping call through to production and beyond. You design how the agent behaves, write the code that connects it to the customer's systems, put it on their phone lines, measure it and keep making it better.}

^{This is not sales engineering and it is not support. You ship code that runs in production, and you own what happens after it ships.}

^{We want someone who already builds agents because they can't help it, and who wants to do it at the hard end: real-time voice, messy enterprise systems, and callers who notice every mistake.}

^{What you'll do}

• ^{Own agents in production, end to end. Scope with the customer; design the conversation, the tools and the guardrails; build in Agent Studio and our ADK; launch; and stay accountable for how it performs.}
• ^{You have built and shipped LLM agents in production, including tool use and evals} 
• ^{Connect agents to the real world. Integrate CRMs, booking engines, payment flows and twenty-year-old APIs in Python. Work across SIP, telephony and contact-centre platforms. Make tool calls fast and reliable in the middle of a live conversation.}
• ^{Debug live calls. When a call goes wrong, work out whether it was the speech recognition, the model, the prompt, an integration or the phone line, and fix the right thing.}
• ^{Run agents like a scientist. Build eval sets from real calls, run A/B tests and let data decide what ships. Containment, resolution and caller experience are your numbers.}
• ^{Close the gap to the platform. When you solve something for one customer, turn it into a reusable tool, template or product change so the next deployment is faster. Work directly with product, research and platform engineering, including the team training Raven, our own LLM built for customer service.}
• ^{Be the technical lead the customer trusts. Run sessions with their engineering, operations and executive teams. Explain trade-offs plainly. Say no when no is the right answer.}
• ^{Work AI-native. You use coding agents and LLM tooling every day, and you help the people around you do the same.}

^{You'll fit if}

• ^{You have shipped LLM-powered agents or conversational systems to real users, and you can show us one.}
• ^{You write strong Python and are at home in production code, APIs and integrations, including codebases you didn't write.}
• ^{You have several years in a technical, customer-facing engineering role: forward deployed, solutions engineering that ships real code, consulting engineering, or a startup where you were the engineer customers called.}
• ^{You own outcomes, not tickets. You have been the person on call when it broke, and you liked it.}
• ^{You can explain a latency budget to an engineer and a containment trade-off to a COO in the same afternoon.}

^{Even better}

• ^{Voice and telephony: SIP, IVR, CCaaS platforms, real-time audio.}
• ^{Evals and experimentation, or an ML background: you have built a test set or run an online experiment.}
• ^{Regulated industries such as healthcare, utilities or financial services.}
• ^{Go or C++, or high-throughput real-time systems.}

^{Why PolyAI}

• ^{Voice is the hard channel, and it's all we do. Real-time, interruptible, noisy, on phone lines built decades ago. Doing it well at enterprise scale is a moat, and you will learn things here you can't learn anywhere else.}
• ^{We own the stack that matters. Raven is our own LLM, trained for customer service inside the agent harness. When the model is the problem, the people who can fix it sit down the corridor.}
• ^{Real production, real stakes. Your agents go live for brands people have heard of, and you will see the impact in their numbers within weeks.}
• ^{Ownership early. Small teams, each engineer owning accounts. The best ideas from the field become the product.}
• ^{London HQ. Founded by conversational-AI researchers from Cambridge, with teams in London and North America.}

^{We offer competitive compensation based on experience, expertise, and the level of responsibility. This role also includes equity, giving you the opportunity to share in the long-term success of the business. The listed expectations reflect what we're hiring for, so we encourage you to review the job description carefully.}

---

^Benefits

^{💰 Participation in the company’s employee share options plan}

^{🏝 Tenure-Based PTO: You will receive 25 holidays when you join and will gain an additional 1 day after 2 years of service, then 1 day each year until capped at 32 holidays}

^{🏡 Flexible working from home policy}

^{🌎 Work from outside of the UK for up to 6 months each year}

^{🫂TELUS Health EAP 24/7 - offers you and your chosen family confidential, judgment-free support for any work, health, or life challenge}

^{🧡 Enhanced parental leave}

^{🚲 Bike2Work scheme}

^{📚 Annual learning and development allowance}

^{🏡 We’re all about making WFH work for you - that’s why we offer a one-off WFH allowance when you join. Offering perks like noise-cancelling headphones or a comfortable desk chair to boost your comfort and focus!}

^{👨‍👩‍👧 Company-funded fertility and family-forming programmes}

^{🌸 Menopause care programme with Maven}

^{🏥 Private healthcare and dental cover, discounts on gym members and relaxation apps, and access to a range of mental health programs}

^{🌎 Sabbatical Program: 5-week paid sabbatical available after 5 years of employment}