William Dennis

Machine Learning Researcher & Computational Modeller
Building robust systems using ML/AI with 3 years of industry experience.

Experience

Energy Modeller

Aurora Energy Research · Oxford, UK
Oct 2024 – Present
  • Applied convex optimisation and maximum likelihood estimation methods to forecasting energy markets.
  • Presented and developed a use case for reinforcement learning to improve the energy pricing methodology.
  • Lead modeller of the Polish, Czech and Slovak energy market model, building features and debugging user errors.
  • Developed a range of tooling in Python using Streamlit for clear data analysis and SQL queries.
  • Applied best software engineering practises with test driven development, git control and agile/scrum team planning.

Data Scientist

JP Morgan · Bristol, UK
Feb 2024 – Jul 2024
  • Lead a team of 4 data scientists to produce in-house quant level research into the development of trading signals derived from millisecond level limit-order-book data.
  • Built tooling for my team to parse millions of rows of raw text data on local machines using JIT.
  • Created baseline models using ARIMA, RNNs and LSTMs on level-1 data.
  • Engineered 50+ features derived from cleaned up level-2 limit order book data.

Signal Processing Engineer

Milbotix · Bristol Robotics Lab, UK
Jun 2023 – Oct 2024
  • Developed a two-stage machine learning pipeline optimised for use on embedded devices using accelerometer filtering and novel Fourier features.
  • Internship extended twice to aid the development of SmartSocks© into a medical grade device.
  • Achieved an error rate of <3bpm, allowing the device to be medical grade.

Intern Hedging Analyst (Risk Team)

Ocean Partners · Maidenhead, UK
Jul 2021 – Jul 2022
  • Mitigated £250m of physical commodity transactions via the London Metal Exchange.
  • Maintained a strong relationship with banks and brokers using responsive and professional communication.
  • Automated a data analysis task for clearing transactions, saving 60 minutes of manual labour daily.

Education

University of Bath

BSc (Hons) Mathematical Sciences
Sep 2019 – July 2023
Scientific Computing (84%) and Machine Learning (91%)
  • Built websites in Python to support startups in Setsquared using Django.
  • 2019/2020 poker society winner.

University of Bristol

MSc Data Science
Sep 2023 – Sep 2024
Distinction (Top 3 of cohort)
  • Deployed distributed cloud architectures on AWS using DynamoDB, SQS, SNS, S3 and EC2 services.
  • Produced a report on migration patterns of people aged 20-25 in the UK utilising data visualisation.
  • Member of Google Developer Student Club and Quantitative Analyst at Bristol Trading Society.

Research & Publications

Publications

William Dennis and James Pope

ICAART 2025 (17th International Conference on Agents and Artificial Intelligence) Feb 2025

Robust Machine Learning for Harsh Environments: A Framework and Evaluation of Key Architectural Choices

Published

William Dennis and James Pope

Lecture Notes in AI, LNAI Dec 2025

Research Projects

Moonboard-Analysis

2025-2026

ML benchmarking framework for climbing route prediction. Designed a reproducible evaluation protocol with 5-fold retrain-per-fold cross-validation, tested 8+ model architectures (MLP, LSTM, CNN, Transformer, Ridge, Perceptron), and established a non-tree leaderboard. Built dataset augmentation pipeline with hold-swap tokenisation and identified/patched data leakage bugs.

PyTorchMLResearchBenchmarking

Robust ML in Harsh Environments

2025

A novel framework in Python with PyTorch for evaluating neural networks in harsh environments. Proposed two novel activation functions that proved to be more robust to SEE related errors.

PyTorchResearchComputer Vision

Classification of Signal Quality from Ankle-Based PPG Signals

2024

Developed novel time-series features for classification of photoplethysmography signal segments. Tested a range of ML architectures and hyperparameters using AUC+ROC metrics. Manuscript prepared but not submitted for publication.

Signal ProcessingTime-SeriesMedical ML

Technical Skills

Cloud & Tools

AWSGoogle CloudDockerGitJIRAAgile

Libraries & Frameworks

PyTorchTensorFlowScikit-learnPandasNumPyOpenCVStreamlitDjango

Languages

Python (10 years)SQLC#RMATLABHaskellTypeScript