RN

AI Engineer · PhD · London

Robin Nicole, PhD

I help teams put AI to work, without getting lost in the jargon. From figuring out what you actually need, to shipping something that quietly runs in production.

See a project I've built

You don't need to be technical to work with me. Explaining things clearly is half the job.

I built an AI assistant that let a customer team handle twice as many questions, without hiring more people.

A story from my work

How I helped a team handle twice as many customer questions

AI customer assistant

The problem
The customer operations team was the bottleneck. Every question from a client required someone on the team to dig through scattered documents, past tickets, and dashboards to find the answer. It was slow, the team couldn't keep up, and hiring more people wasn't going to fix it.
What I built
I built an AI assistant that reads through the company's internal knowledge and answers questions accurately, with links back to the original source so anyone can double-check. I also set up the safety rails (to keep costs under control and answers appropriate) and a way to measure whether quality held up as the content kept changing.
What changed
The team could handle twice as many customer questions without adding headcount. Response times dropped from minutes of searching to seconds. It's still running in production today, on-budget and quietly reliable.

Common questions

You might be wondering

Q & A
I'm not technical, can we still work together?
Yes. Most of my work involves translating between what a business actually needs and what's technically possible. You don't need to know acronyms, frameworks, or anything about how AI works under the hood. Part of my job is explaining all of that in language that makes sense for your world.
Q & A
How do I know if my problem is even an AI problem?
Honestly, many aren't. A lot of what gets labelled an 'AI project' turns out to need better dashboards, cleaner data, or straightforward automation. Tell me what you're trying to achieve and I'll tell you whether AI is the right tool, or whether something simpler would serve you better.
Q & A
What if our data is messy or incomplete?
Welcome to almost every company I've worked with. Part of the first few weeks of any engagement is looking at what you actually have, figuring out what can be done with it today, and flagging what needs cleaning up before we go further. You don't need a perfect data set to get started.
Q & A
How does an engagement usually start?
A short no-commitment call so I can understand what you're working on and what you've already tried. If I think I can genuinely help, I send a proposal with a clear scope, timeline, and price. No surprises. If I don't think I'm the right fit, I'll tell you, and I'll usually know someone who is.

How I can help

What I can build for you

Service

AI assistants & smart search on your own data

Think a chatbot or search tool that reads through your company's documents, knowledge base, or product catalog, and answers questions accurately, with links back to the source so everyone can verify. I handle the setup, the guardrails that keep it safe and on-budget, and the follow-up so quality stays high as your content changes.

Typical engagement: 2 weeks to 3 months

Service

Forecasting and pricing, powered by your data

Predicting demand, setting prices, and optimizing operations using the data you already have. I'll help you figure out what's achievable with what you have today, then build it into a system your team can rely on for real decisions.

Typical engagement: 4 weeks to several months

Service

Getting AI systems running reliably

Many companies have AI prototypes that never made it into real use. I help close that gap, turning experiments into software that runs dependably, stays monitored, and fits how your team actually works.

Retainer or a scoped platform build

Where I've been

The work that got me here

Track record

Seven years across retail, adtech, media, and quantitative finance. Production ML, not prototypes.

  1. Oct 2024, Present

    Senior ML Engineer · Kingfisher plc

    Building the forecasting and pricing systems behind a major European retailer. Turning raw sales data into decisions teams can trust every day.

  2. Jan 2022, Sept 2024

    Senior Applied Scientist · Sojern

    Machine learning for a travel advertising company. Built the AI assistant that doubled the customer team's capacity, plus the automation that kept models running without babysitting.

  3. Sept 2019, Jan 2022

    Senior Data Scientist · Reach plc

    Language and text systems for one of the UK's largest news publishers. Content classification, real-time pipelines, and a testing framework editors could actually use.

  4. May 2018, Sept 2019

    Quantitative Analyst · Gambit Research

    Predictive models for financial markets. Learned to ship code that had to be right under real-world pressure.

About me

A bit about me

I'm Robin Nicole, a senior ML engineer with a PhD in Applied Mathematics from King's College London. I've spent 7+ years building production ML systems across retail, adtech, media, and quantitative finance.

Most of what I do boils down to two things: figuring out what's actually possible with the data and constraints you already have, and then building something that quietly works in the background for years. I've shipped demand forecasting pipelines that price products across European retail, and AI assistants that doubled how many customer questions a team could handle in a day.

What I care most about is making this stuff feel less intimidating. You shouldn't need an engineering degree to know what's happening with your own data. If we work together, expect weekly updates in plain language, honest answers when something is harder than I thought, and a working system at the end you can hand to your team.

BlogLinkedInGitHub

Published work

What I work with

What I work with
AI assistants & chatbots, Search over documents, Forecasting & pricing, Automation pipelines
Tools & languages
Python, SQL, TensorFlow, PyTorch, scikit-learn
Where it runs
Google Cloud (VertexAI, Kubeflow), AWS, Docker, Your own servers

Say hello

Let's talk about your project

Have an AI or data problem that's been on your team's plate for a while? I'd genuinely love to hear about it.

Tell me what you're trying to do and what you've already tried. Don't worry about using the right words. I'll reply within a week, and if I'm not the right fit I'll happily point you toward someone who is.

Get in touchor just email robin.nicole.m@gmail.com