Forward Deployed Engineer (Anthropic)

About Digital Futures

Digital Futures is an AI services and software company that specialises in accelerating AI transformation and enabling AI governance for enterprise and government.

The Role

Digital Futures is mobilising a forward-deployed team to work alongside Anthropic’s Applied AI team, embedding directly with a strategic client to accelerate enterprise adoption of Claude. You will collaborate closely with business and technology teams to ship production AI applications that solve real-world problems, working hands-on with users and delivery teams inside the client’s environment and iterating quickly based on their feedback.

This is a hands-on build role, not an advisory one. Using Claude Code and the Claude API, you will design, build and evaluate agentic systems that make sense of complex workflows and legacy data, and own the quality of what they produce. Governance is integrated from day one, with every output evidenced and validated by domain specialists, and you will capture repeatable deployment patterns so that capability can be scaled and transferred to client teams. We expect our forward-deployed engineers to operate autonomously, thrive under ambiguity and maintain high standards of safety and reliability in regulated environments.

Key Responsibilities

Agent design and build

• - Design and build the five discovery agents (requirements, estate, lineage, data quality and documentation) using Claude Code and the Claude API, delivering artefacts such as MCP servers, sub-agents and agent skills under read-only, governed and logged access.
• - Engineer agents that trace transformation chains through SAS, SQL, Hive, Spark and shell code, infer field semantics from code, naming, constraints and usage, and detect where source, derivation or meaning changed over time.
• - Contribute to joint architecture reviews with the bank’s data platform and security teams and Anthropic’s Applied AI team, covering agent design and the evaluation plan, and present early test results at architecture checkpoints.

Evaluation and quality

• - Build the evaluation harness: ground-truth sets from specialist-validated entries, calibrated confidence scoring, failure analysis and regression testing as corrections accumulate.
• - Ensure every generated entry carries evidence links, provenance, a confidence score and a named reviewer, written to the evidence store and audit log.
• - Use data profiles (null rates, value distributions by period) to test inferences against the data itself, reducing entries that are plausible but wrong.
• - Close the loop with weekly specialist validation sessions, so that corrections feed back into the agents’ working context and accuracy improves throughout the engagement.

Reuse and capability transfer

• - Package the agents, evaluation harness and playbook as a reusable toolkit that applies to further entities and data stores with minimal change, codifying repeatable deployment patterns with Anthropic’s Applied AI team.
• - Pair with the bank’s engineers, transferring capability so that future data stores can be addressed by bank teams with Digital Futures support rather than dependence.

Experience and Profile

• - Production experience building agentic systems with large language models, with demonstrable depth in Claude Code and the Claude API, including tool use, the Model Context Protocol (MCP), sub-agents and agent skills.
• - Evaluation of LLM output where accuracy is material: harness design, ground-truth construction, confidence calibration and systematic failure analysis.
• - Experience building systems whose output is validated, evidenced knowledge rather than code, with feedback loops that make them more accurate as specialists review them.
• - Delivery experience in enterprise financial services, ideally a Tier 1 bank, including working within security tollgates, change control and third-party access constraints.
• - Fluency reading legacy data code (SAS, SQL, HiveQL, PySpark, shell and scheduler scripts) well enough to reason about derivation logic and its history.
• - Python to production standard and a disciplined Git-based workflow.
• - Experience with data lineage tooling, knowledge graphs or graph databases is desirable.
• - A previous forward-deployed or client-embedded engineering role, with the autonomy to thrive under ambiguity in complex organisations.
• - The ability to explain design choices, limitations and residual risk clearly to model risk, security and senior stakeholders.

Benefits

We offer a competitive salary. Our benefits package is designed to support your wellbeing and development. This includes:

• - Group Life Assurance
• - Virtual GP and Online Physiotherapy
• - Confidential Helpline/Employee Assistance Programme
• - Wellbeing Studio & Digital Gym
• - Discounted shopping cards, e-vouchers & cashback offers
• - Percentage discounts & savings on brands online & in store
• - Statutory pension contribution - salary exchange scheme
• - 30 days annual leave, including UK Bank Holidays

Why Join Us

Digital Futures consultants deliver solutions across the full data and AI transformation lifecycle, from strategy, analytics and engineering through to adoption, governance and assurance.

We’re focused on creating opportunity in the age of AI. Joining Digital Futures means joining an organisation making a real impact in delivering an AI-first future, at whatever stage of your career.