Software Engineer / AI Engineer

53.38° N, 1.47° W — Sheffield, UK

Aaron Royan

I build software that keeps real operations moving — and bring an MSc in AI to the problems that need it.

Now
Junior Developer, A Shade Greener
Before
AI Model Engineer, The King’s Trust
Full Stack Developer, Loyalytics.AI
Education
MSc Artificial Intelligence, University of Sussex

Fig. 00 — 2-opt route solver, running live 0 stops · 0% shorter

Open to conversations

Scroll

(01)

Selected work

2021 – 2026
Six projects

  1. 01 Warehouse Management System A Shade Greener Product engineering 2026
  2. 02 Engineer Route Optimisation A Shade Greener Optimisation 2026
  3. 03 Roofing Job Management A Shade Greener Field operations 2026
  4. 04 Cognitive State Detection The King’s Trust Applied ML · Research 2025
  5. 05 LLM Prompt Recovery University of Sussex NLP · LLMs 2024
  6. 06 ML Launch Programme Loyalytics.AI ML · APIs · Cloud 2022

Warehouse Management System

Fig. 01 — Pick-path schematicLive

A full-scale warehouse management system, built from scratch for the company’s core workflows.

Warehouses run on accuracy: knowing what’s in stock, where it lives and what needs to move next. I built A Shade Greener’s warehouse management system from a blank repository to handle those core workflows.

Engineer Route Optimisation

Fig. 02 — Travel-cost matrix → optimised routeLive

Route-matrix optimisation that plans engineers’ routes more efficiently.

Every extra mile between jobs is time an engineer isn’t on site. The approach: build a matrix of travel costs between every pair of stops, then search for the visiting order that keeps the total as low as possible — the same idea running live at the top of this page.

Roofing Job Management

Fig. 03 — Job board schematicLive

A job management system for the off-site roofing team.

Software for people whose workplace is a roof, not a desk. Alongside it I maintain and extend the rest of the company’s software — fixes, new features and continuous improvements — on k3s clusters across development and deployment.

Cognitive State Detection

Fig. 04 — EEG artifact removal (illustrative signal)Live

A wearable that reads mental fatigue from brain signals.

I designed a wearable bio-sensing system combining EEG (the brain’s electrical activity) and fNIRS (blood oxygenation) for mental-state analysis — then built the pipeline that turns noisy raw signals into predictions: artifact removal, feature extraction and machine-learning fatigue classification, tested on real-world data.

  • +42%Data accuracy
  • +35%Classification accuracy
  • −60%Processing time

LLM Prompt Recovery

Fig. 05 — Original + rewrite → recovered promptLive

Working backwards from a language model’s output to the prompt that produced it.

Given an original text and a version rewritten by a large language model, can you recover the instruction it was given? Prompt recovery was the specialist focus of my MSc in Artificial Intelligence and Adaptive Systems: reverse-engineering intent from text.

ML Launch Programme

Fig. 06 — Model served through scalable APIsLive

A machine-learning launch programme that lifted client profits by 7%.

I led the design and deployment of a new launch programme, integrating machine-learning models with scalable APIs. Alongside it I built full-stack web apps for real-time data analysis and decision-making, working in cross-functional agile teams on tight timelines.

  • +7%Client profit
  • Best Team LaunchCompany honour

(02)

Experience

2021 – Now

  1. Feb 2026 – Now

    Junior Developer

    A Shade Greener · Sheffield

    • Built a full-scale warehouse management system from scratch for core workflows.
    • Built a job management system for the off-site roofing team.
    • Implemented route-matrix optimisation to plan engineer routes more efficiently.
    • Maintain and develop company software through fixes and continuous enhancements, working with k3s clusters across development and deployment.

    ReactTypeScriptk3s

  2. Mar – Dec 2025

    AI Model Engineer & Researcher

    The King’s Trust

    • Designed a wearable EEG + fNIRS system to capture brain activity for mental-state analysis, improving data accuracy by 42%.
    • Built the pipeline for artifact removal, feature extraction and ML fatigue classification, improving classification accuracy by 35%.
    • Developed and tested the system on real-world data, reducing processing time by 60%.

    PythonMNEscikit-learnOpenBCI

  3. Oct 2021 – Jun 2022

    Full Stack Developer

    Loyalytics.AI · Bangalore

    • Led the design and deployment of a new launch programme, integrating ML models and scalable APIs to boost client profits by 7%.
    • Built scalable full-stack web apps for real-time data analysis and decision-making, with TypeScript backends on AWS, GCP and Azure.
    • Delivered with cross-functional agile teams on tight timelines using Python and Node.js.

    TypeScriptNode.jsPythonAWSGCPAzure

  4. 2023 – 2026

    Alongside: professional kitchens

    Miller & Carter · Tonkotsu

    Chef, then Junior Sous Chef, through my MSc and research. Led teams through busy service and redesigned kitchen workflows to lift profit by 8% and cut food waste by 20%. Voted Player’s Player and named Employee of the Quarter.

(03)

Capabilities

Tools I reach for

Languages 05

  • TypeScript
  • JavaScript
  • Python
  • Java
  • C++

Product 05

  • React
  • Node.js
  • REST APIs
  • ASP.NET Web API
  • Client–server design

AI / ML 05

  • PyTorch
  • scikit-learn
  • BERT & NLP
  • MNE · OpenBCI
  • Model training

Infrastructure 05

  • Docker
  • k3s / Kubernetes
  • Microsoft Azure
  • AWS · GCP
  • DevOps practices

Data & practice 05

  • MongoDB
  • Database administration
  • NumPy
  • Software testing
  • Agile delivery

(04)

About

Sheffield, UK

I’m a software engineer in Sheffield who enjoys the unglamorous systems real operations run on. My path is a little unusual — an MSc in AI, research turning brain signals into fatigue predictions, and time in professional kitchens, which is still the best training I know for staying calm when production is on fire.

Education

  • MSc Artificial Intelligence & Adaptive Systems

    University of Sussex · 2023 – 24

    Specialist focus: LLM prompt recovery

  • Bachelor of Computer Applications

    St. Joseph’s University, Bangalore · 2019 – 22

    Student council; built full-stack websites for college departments

Honours

  • Best Team Launch

    Loyalytics.AI

  • Recognition award, Neerathon platform

    St. Joseph’s University

  • Player’s Player & Employee of the Quarter

    Miller & Carter

(05)

Archive

Earlier work
from university

  1. 2022 Neerathon Official registration website for St. Joseph’s annual Neerathon marathon, built with my team. Won a recognition award. Live ↗Code ↗
  2. 2022 Beer Library Nuxt web app for searching and filtering beers through the Punk API. Code ↗
  3. 2022 Mars Invaders Arcade shooter written in Python with pygame, inspired by Space Invaders. Code ↗
  4. 2022 Alexa A minimal voice assistant in about sixty lines of Python. Code ↗