About Me
Table of Contents
Who am I? #
I’m a senior data scientist and AI engineer, with a background in cognitive psychology.
I’m on the job market #
Are you trying to build products powered by the latest generation of AI models? Or struggling to figure out how to evaluate the performance and safety of such a product? Or need to do some good-old-fashioned machine learning, causal inference, or even data viz?
If yes, and especially if you’re working in the psychology/mental health space,
I’d lke to chat to you. Get in touch, at{firstname}.{lastname}@gmail.com
Where have I been? #
Senior Data Scientist | Unmind | B2B2C | Jan 2023 - Present
I was the sole data scientist at Unmind when we pivoted to generative AI, and developed the companies AI-powered products from scratch. I designed, built, and productionised Nova, (an AI wellbeing coach chatbot), and the AI practitioner matching feature, which matches users to human therapists/coaches.
I also led work on our evaluation of the usefulness and safety of these features, working closely with clinicial psychologists, product, and commercial teams to build evaluation suites, and enhanced analytics pipelines, including LLM-powered pipelines similar to Anthropic’s Clio.
Of the two, I’m proudest of the evaluations. Building with AI is relatively easy, but figuring out how to be confident that it is safe and effective, and demonstrating this to risk-averse large enterprise clients to build confidence in the product, is hard.
I also supported AI/data/statistics initiatives across the business, including running the internal AI Forum.
Senior Research Scientist | Unmind | B2B2C | Jan 2021 - Dec 2022
Previously, I was in the Science team at Unmind, where I spent most of my time writing, psychometrically validating, benchmarking, and building data pipelines for the Wellbeing Tracker and Organisational Health Benchmark platforms.
Postdoctoral Researcher | UCL Institute of Cognitive Neuroscience | Neuroscience & Mental Health Lab | 2020 - 2021
In my final postdoctoral role, during COVID, I worked with the Neuroscience & Mental Health Lab (PI: Jonathan Roiser) developing a suite of web-based experiments for studying the relationship between different dimensions of mental health symptoms and the individual differences in the computations underlying decision-making. I also trained staff and students in the lab, including the MSc/MRes cohort, on web development, enabling them to modify the experiments I built and create their own for future research.
See this post for teaching materials created as part of this work.
Postdoctoral Researcher | UCL Institute of Cognitive Neuroscience | Action & Body Lab | 2017 - 2020
I was a postdoc in the UCL Action & Body lab (PI: Patrick Haggard) where I researched the neural basis of self-initiated actions. using EEG, computational modelling, and fMRI.
During this time, I also created Evidently, a python toolkit for simulating from the kinds of evidence accumulation models used in this research.
Postdoctoral Researcher | UoL | Institute of Philosophy | 2016 - 2017
For my first postdoctoral position, I worked in the Institute of Philosophy at the University of London School of Advanced Study. In this highly interdisciplinary department, I worked with philosophers and neuroscientists (Nicholas Shea, Chris Frith, Ophelia Deroy, Barry Smith, Merle Fairhurst) using experimental psychology techniques (masked priming, eyetracking, chronometry) to answer questions about consciousness, metacognition, and multisensory processing. I also worked on a number of high-profile public outreach projects during and after this time, including work at Tate Modern (2017, 2018), Cheltenham Science Festival, BBC Radio, and the SAS Living Literature festival (where we had a lot fun using electromyography and electrical nerve stimulation to create an interactive exhibition called The Spark of Life).
It was also at this point that I started teaching workshops on statistical methods. You can find a lot of relevant stuff from these workshops in my blog.
PhD Researcher | QUB | School of Psychology | 2012 - 2015
I did my PhD at Queens University Belfast (supervisors: Aidan Feeney, Jonathan Rolison), using mouse cursor trajectories as a tool for studying conflict in high-level reasoning and decision-making.
What Can I Do? #
AI/ML #
- LLMs & Foundation Models
- OpenAI, Bedrock, HuggingFace, and the rest
- Fine-tuning, Embeddings, RAG
- Evaluation
Engineering #
- Favourite languages: Python, R
- Backend: Flask, FastAPI, SQLAlchemy, etc.
- Frontend Development
- DevOps, CI/CD, and AWS
Data #
- Data engineering: Dagster, dbt, etc.
- Analytics, including LLM-powered pipelines
- Statistics: Causal Inference, Experimentation, Bayesian Methods, Multilevel models
Domain Expertise #
- AI Product Development
- AI Evaluation
- Cross-functional team leadership
- Taking products from zero to one
- I can speak to commercial, clients, and the public, not just techies
- Science: Psychology, Cognitive Science, Neuroscience, Decision-Making, Statistics
There's no articles to list here yet.