Forward Deployed Engineer

About the role

FDEs are how Duvo brings AI to life inside our customers. You sit at the front line, map their workflows on the Duvo platform, improve the underlying processes, and turn the result into running agents that deliver tangible business impact. Scoping, build, debug, handover, expansion. You own the customer outcome end to end and you own the relationship that goes with it.

You won't write the core product. You will write the glue: integration scripts, agent configurations, data transformations, prototypes, whatever it takes to get an agent running against a customer's actual systems. You'll spend real time inside their stack and inside their team.

If you like the messy middle between AI, process, and customer, this is the role.

You will be deployed directly with our clients, mostly in onsite/hybrid mode of work, you will travel a lot.

Location. This role is UK-based. You'll need to be living in the UK, or relocating there before you start, and able to travel to client sites from there.

What you'll do

Embed with customers from kickoff through go-live and beyond

Map customer workflows directly on the Duvo platform and decide what's worth automating, what's worth fixing first, and what's blocked

Improve the underlying processes where it matters. Don't automate broken flows, redesign them

Translate business logic into agent behavior: actions, decision rules, exception paths, fallbacks

Configure and ship agents against real customer systems (ERPs, portals, spreadsheets, email, custom tools)

Write integration code, scripts, and prototypes when the situation calls for it

Debug live agents in production environments and iterate quickly

Build and own the working relationship with customer stakeholders, from operators to sponsors

Run working sessions, hold your own with technical counterparts, and earn the right to push back

Make AI real for the customer. Move them from pilot energy to measurable business outcomes

Capture reusable patterns, templates, and runbooks so the next deployment is faster

Send recurring workflow patterns back to Product and Engineering as real signal

What success looks like

Multiple agents in production across clients, with measurable business impact

Customers point to specific KPIs that moved because of your work

Customers treat you as a trusted operator and want you in the room

Process improvements you drove are visible in the customer's day-to-day, not just in the agent

Transformation Managers want you on their hardest accounts

Accounts you own expand because the relationship and the results are both there

Each deployment is faster than the last because you've left behind reusable assets

Who you are

Experience

5+ years in roles where you've shipped against real customer environments. We're prioritizing AI-native candidates: people who have been building, deploying, or operating AI and agentic systems in customer-facing settings, not adjacent to them.

Likely backgrounds:

Forward-deployed engineering or solutions engineering at an AI-first company

Applied AI / ML engineering with direct customer exposure

Technical consulting on AI or automation projects where you owned delivery and the client relationship

Operations or product roles where you've built and shipped AI-powered tooling against messy real-world systems

Retail, e-commerce, or CPG background is a strong plus. Fluency in any operational domain (supply chain, procurement, category management, merchandising, finance ops, supplier management, warehousing) helps you ramp faster on customer accounts. What matters most is that you can pick up a new business quickly, see how to improve it, and build something with AI that actually works.

Skills

Hands-on with modern AI tooling: LLMs, agent frameworks, prompt engineering, evals, MCP, retrieval

Strong systems thinking and engineering judgment

Process design instinct. You can spot waste, redundancy, and bad handoffs and propose something better

Comfortable in code-adjacent territory: SQL, Python or JS scripting, APIs, JSON, webhooks, light data work