Industry work

I ship insurance products for a major UK brand, and I build the AI tooling that makes my team faster.

I am a Senior Experience Engineer at Publicis Sapient, on the account for a leading UK roadside assistance and insurance provider. I am the lead Angular developer there, on the customer journeys for insurance, savings and loans that millions of people use.

Over the last two years the job has grown past the frontend. I take AI from an internal experiment to something the client backs, and I build the agents and automations that change how the team works day to day. I care as much about making other engineers faster as about my own output.

Employer
Publicis Sapient
Role
Senior Experience Engineer
Client
UK roadside assistance and insurance
There since
2021

What I bring to a team

Production delivery at scale

Lead Angular developer across insurance and financial services journeys for a major UK brand.

AI from demo to funded work

Client-facing AI prototypes that turned into a real project stream.

Engineering productivity

AI-assisted testing and debugging, measured: 10 to 20% of my own story time saved, 15 to 20% less effort per story across the team.

Force multiplier

Mentoring, knowledge-transfer sessions and automations that raise the whole team's output, not only mine.

Case studies

The client is not named, by agreement. The numbers are measured, not estimated.

Flagship

From AI experiment to a client-backed project

AI demos that read B2B contracts for call-centre agents, and moved AI on the account into a formal project stream.

  • Angular
  • LLM APIs
  • Prompt engineering
  • RAG
  • Agentic workflows
  • Batch pipelines

Call-centre agents take roadside breakdown calls for business customers. What cover applies is buried in long, unstructured B2B contracts, and the agent has to find it while the caller waits.

What I did

  • Designed and built AI demos that pull the cover terms out of those contracts and sort them, so an agent can give the caller a quick, correct answer.
  • Showed the working demos to client leadership as part of a wider plan to rebuild the help platform.

Outcome

  • Senior client stakeholders backed the work, and AI on the account moved from internal experiments to a client-backed initiative with its own project stream.
  • Account leadership credits it with changing how the client sees us: as a partner that brings ideas to the roadmap, not only one that delivers.

Related builds

B2B fault-matrix assistant
A proof of concept that builds a rules matrix from contract text, scores its recommendations against data filled in by subject experts, runs policies in batches and writes reports. Reviewed regularly by account leadership as a step toward online breakdown reporting for business customers.
Agentic proof of concept
An agent that runs multi-step tasks on its own for client use cases, to show where agentic workflows fit.
Figma to code
Demos to client leadership of AI-assisted code generation and scaffolding from designs, and how much manual work it removes.

A development agent in a stack I had never used

Picked up .NET from scratch to help build a development agent for the account's always-on engineering model.

  • .NET
  • Agentic development workflows
  • Microsoft Copilot

The account is moving to an agent-enabled way of engineering, where AI agents take on parts of the development workflow around the clock.

What I did

  • With no prior .NET experience, learned the stack and helped build a .NET development agent on a short turnaround.
  • It feeds directly into the account's always-on engineering model.

Outcome

  • New account leadership recognised the work formally within weeks of taking charge.
  • It is the pattern I am proudest of: walk into an unfamiliar area, learn fast, and turn the learning into something that works.

Making the whole team faster with AI

Ran the GitHub Copilot proof of concept and the enablement that followed, with measured gains.

  • GitHub Copilot
  • Microsoft Copilot
  • Dependabot
  • Angular
  • TypeScript

The goal was never AI for its own sake. It was to put AI into everyday development, so it stopped being an innovation exercise and became how the team ships.

What I did

  • Ran the GitHub Copilot proof of concept on the engagement: built the test-generation and debugging use cases, wrote the developer feedback report, and ran demos so every engineer could use it.
  • Recommended and drove AI-assisted unit test generation, and Dependabot automation that raises pull requests for dependency alerts on its own.
  • Built an automated workflow on Microsoft Copilot with limited resources, and documented it so others could build on it.

Outcome

  • 10 to 20% of my own sprint-story time saved, mostly on unit tests and debugging.
  • 15 to 20% less unit-testing and debugging effort per story across the team after enablement, with a per-sprint velocity improvement reported by engineering leadership.
  • The AI-assisted unit testing recommendation showed roughly 50% less testing effort, about 15% of overall development time saved.

Lead Angular engineering for a major UK insurer

The core insurance and financial services journeys, the Angular 19 upgrade, and the standards around them.

  • Angular (through 19)
  • TypeScript
  • NgRx
  • Unleash
  • CI pipelines
  • WCAG
  • Lighthouse
  • Dependabot

Insurance and financial services front ends for a brand with millions of customers, in a regulated industry where a bad release is expensive.

What I did

  • Lead Angular developer on the account. Product owners and business analysts bring me into requirement discussions early.
  • Delivered savings and loans journeys, auto login, aggregator integrations, van insurance, multi-car discounts, eligibility checks, analytics and monitoring, and feature-flag rollouts on Unleash.
  • Supported the Angular 19 upgrades across the insurance repositories, NPM vulnerability management, and crash-rate work on a key app surface.
  • Pushed frontend standards: shared UI libraries, automated accessibility and Lighthouse checks in CI, simpler routing.
  • Own critical releases and live issues, including rollbacks and DevOps bridges, and keep leadership informed while they happen.

Outcome

  • Seen on the account as the person who sets the bar for frontend quality, and trusted with the releases that cannot go wrong.

Leadership and mentoring

The part that does not show up in a commit log.

  • Run knowledge-transfer sessions across the account, from AI-assisted development and Copilot adoption to domain and codebase walkthroughs, so knowledge does not sit with one person.
  • Mentor engineers through onboarding, pairing and code review until they own feature delivery that used to need senior oversight.
  • Unblock teammates wherever needed: production bugs, onboarding, new joiners on unfamiliar journeys.
  • Help with hiring: technical screening, interviews and referrals for frontend roles.
  • Built two team automations that are now part of how the team runs: a weekly engineering status report published to stakeholders, and a work-anniversary appreciation workflow.

Recognition

  1. 2023 and 2025
    Recognised twice by Publicis Sapient's CEO

    For hackathon work, fresh ideas and a learning mindset.

  2. 2023 to 2026
    Four global hackathons in a row

    Frontend build, architecture, product roadmap and product-and-engineering roles. Concepts included an AI learning platform simulation, an AI employee engagement hub and a decision acceleration platform.

  3. 2026
    The account's AI catalyst

    Recognised by engineering leadership for moving the team from exploring AI to using it in everyday engineering, and for documenting the work so others can build on it.

  4. 2026
    Recognised by new account leadership

    For the .NET development agent: a new stack, learned fast, shipped.

  5. 2026
    Fault-matrix proof of concept named a strategic initiative

    For its potential to open up online breakdown reporting for business customers.

  6. Ongoing
    Peer recognition

    Through the account's monthly recognition programme, for a proactive attitude and the Copilot test-generation work.

Domains and training

  • Insurance: motor, van, breakdown and roadside
  • Financial services: savings, loans, aggregators
  • Advanced Prompt Engineer
  • Advanced Prompting in GPT-4
  • Prompt Engineering Fundamentals
  • Prompt Engineering for Engineers
  • Generative AI programme with LangChain (45 hours)
  • Gen AI Academy Level 2
  • Sapient Slingshot AI coding programme
  • PL.AI
  • Engineering and Data certification
  • NFR Web Specialist craft programme (46 hours)
  • NFR Performance and Accessibility track

By day I ship insurance products used by millions and build the AI tooling that makes my team faster. The rest of this site is what I build after hours.

Send a messageCVWhat I build after hours