Zach Rodgers
Graduate Data Scientist at Afternoon. First Class AI & Data Science.
Applied LLM and data science at Afternoon, an AI-first operating model for financial advice firms. First Class AI & Data Science, 2026.
About
I build applied LLM systems at Afternoon, an AI-first operating model for financial advice firms, where I work on retrieval, grounding, and evaluation. Before this, First Class Honours in AI & Data Science from Glasgow Caledonian and a research project on deep learning for EEG signals.
Honours Project
My honours project involved designing and implementing seven temporal deep learning architectures for EEG-based digit classification, including two novel hybrid architectures of my own design, and running a controlled comparison across tasks, devices, and input representations to establish where the real performance ceiling lies.
I co-directed Paydit, a fintech startup enabling small businesses and charities to accept payments via QR code, where I contributed backend development in Java and Spring Boot.
I like working close to the product, where the modelling meets a real user.
Before tech, I spent around five years in professional kitchens, most recently running the larder section at a five-star hotel. It left me as comfortable under pressure in a production incident as on a Saturday dinner service.
7
DL Architectures Benchmarked
99.22%
Best Classification Accuracy
RAG + Evals
Focus at Afternoon
Selected work
A systematic comparison of seven temporal deep learning architectures for EEG-based digit classification on the MindBigData corpus - EEGNet, BiLSTM, 1D CNN, TCN, TCAN, and two novel hybrid architectures (EEG-TCN and EEG-TCAN) designed and implemented as original contributions. Models were evaluated across a 10-class digit recognition task and a binary digit detection task, using both raw time-series and bandpower feature representations under identical preprocessing and splits.
Supervised by Dr Yan Zhang, Glasgow Caledonian University
AI Integration & RAG — QikServe
Led AI integration strategy across QikServe's hospitality SaaS suite as Product Intern. Built proof-of-concept programs including a retrieval-augmented (RAG) chatbot using LangChain and GPT, and authored a 26,000-word AI strategy wiki spanning four initiatives. QikServe was subsequently acquired by Access Group.
Place My Mise
Full-stack web application for visualising and structuring kitchen mise en place workflows as interactive graphs. Built end-to-end, from data modelling to React frontend, demonstrating product thinking, graph data structures, and clean software delivery outside of pure ML contexts.
Human Activity Recognition (1D CNN)
TensorFlow implementation of a 1D Convolutional Neural Network on the UCI Human Activity Recognition dataset. Includes TensorBoard integration for training visualisation and a full technical report written in LaTeX, covering architecture decisions, hyperparameter tuning, and evaluation methodology.
Technical toolkit
Languages
ML & Data
Backend & Infra
Frontend
Practices
Tools
Where I've worked
Graduate Data Scientist
Afternoon ↗Applied LLM and data science at a fintech building AI tooling for financial advisers. Focused on making model output trustworthy enough to rely on: grounding, retrieval, and evaluation.
Director
Paydit ↗Fintech startup enabling small businesses and charities to accept payments via QR code, eliminating card machine fees. Operates through a partnership with Lloyds Bank. Contributed backend development including account hierarchy design and multi-account access features.
Product Intern (AI & Data Analytics)
QikServe ↗Led AI integration strategy across QikServe's hospitality SaaS product suite. Conducted internal research and stakeholder interviews, produced a 26,000-word strategy wiki, developed proof of concept programs, and delivered roadmaps for four AI initiatives - all ahead of schedule. QikServe was subsequently acquired by Access Group.
Demi Chef de Partie
Sheraton Grand Hotel & Spa EdinburghLarder section management in a five-star hotel kitchen. Part-time alongside studies and Paydit.
Earlier hospitality roles on LinkedIn
Education
Glasgow Caledonian University
BSc (Hons) Artificial Intelligence and Data Science
Honours Project
Deep learning for EEG-based digit classification using seven temporal architectures on the MindBigData dataset. Supervised by Dr Yan Zhang.
What people say
“His ability to gather information, process it and deliver outputs that can be consumed up and down the org is awesome. What I really valued was he never overly invested in an idea and could adjust, pivot, or drop an idea if that's what the data was telling us. In our world, THAT is powerful.”
Chris Timms
Head of Solutions · QikServe
Let's build
something smart.
Graduate Data Scientist at Afternoon. Always open to connecting with people building interesting things in tech - get in touch.
Prefer copy/paste? Email me directly at zachjrodgers@gmail.com