AI Forward Deployed Engineer

Company Description WealthTech Pros is a UK wealth-management technology consultancy. We help wealth management firms unlock and unify data across their estate to drive competitive advantage, delivering native Agentic AI architectures that sit above existing platforms without replacing core systems. The role We're hiring an AI Forward Deployed Engineer to embed on-site in Edinburgh with a leading UK wealth management client. Working closely within our team, you will turn an agreed AI business case into shipped, measurable value, sprint by sprint. This isn't advisory. By the time you arrive, the target process, data lineage and ROI model are already agreed. Your job is to build it using AI models via CLI, paired with the client's own Product Owners and AI engineer, in two-week sprints, against a costed KPI baseline. What you'll do * Confirm current state onsite: the stack, the workflows, the real pain points. * Turn a ranked shortlist of opportunities into a live delivery backlog. * Define measurable success criteria before build: cost-to-serve reduction, volume, error rate. * Design testable, automatable validation for every initiative, so correctness is checked automatically, sprint over sprint. This is central to the role. * Run paired two-week sprints with the client engineer, who builds inside their own estate under your technical direction. * QA-gate every work item: code and agent-behaviour review, zero-trust data handling, regression against the automated suite. * Run your agent fleet like a workforce: track run-cost per work item, retire or retrain agents that drift. * Prove and report the costed delta each sprint. What we're looking for * Production-grade software engineering: you review and improve, not just prescribe. * Strong test design and automation instincts. Building for verifiability is the job. * Depth in LLM and agentic tooling: Claude, MCP architectures, RAG and agent orchestration. Claude Code and MCP workflows are how we work by default, not an add-on. * Comfortable in regulated, in-network / no-egress environments (FCA, PRA, UK GDPR). * Data literate: read a lineage map and a cost-to-serve model, turn it into a sprint backlog. * Prior forward-deployed or embedded delivery experience (Palantir FDE-style, or boutique AI/data consultancy). * A teaching mindset: you uplift the client's engineer rather than black-boxing the work. * Happy to be based onsite in Edinburgh for the foreseeable term. First 90 days 30: Onsite and trusted. Current state confirmed. First sprint shipped with automated validation designed in. 60: Two initiatives live with eval suites running. Pairing cadence established. 90: First verified, costed saving on the ledger. Client engineer executing work items with reduced oversight. How this differs from a generic AI consultant We embed and ship every two weeks. We work with a named client engineer, not for the client. Success is costed KPI movement, not workshops run. The engagement is the build. Compensation: Competitive + equity Reports to: Rob Duncan, CTO \& Chief AI Officer Interested? Message Rob Duncan or Dave Howard directly, or apply via this post.