Tony

Analysed ∼1000 high-dimensional text observations (patent claims/abstracts) using embedding-based representations to extract latent innovation structure. Applied unsupervised learning (clustering on embedding spaces) and LLM-based summarisation to convert noisy, unstructured data into interpretable signals. Built a Python end-to-end pipeline and Streamlit interface to explore clusters, metadata, and distributions, demonstrating scalable ML-driven discovery.

Data Science Micro-Intern - AIM Reply AI

(2025-12)

• Analysed ∼1000 high-dimensional text observations (patent claims/abstracts) using embedding-based representations to extract latent innovation structure.
• Applied unsupervised learning (clustering on embedding spaces) and LLM-based summarisation to convert noisy, unstructured data into interpretable signals.
• Built a Python end-to-end pipeline and Streamlit interface to explore clusters, metadata, and distributions, demonstrating scalable ML-driven discovery.

Simulations Engineer - Formula Student Electric

(2024-10 - 2025-06)

• Built a race-line optimisation pipeline: Neural Network surrogate approximating MATLAB fmincon; 95% faster inference with ≤ 2% lap-time error.
• Developed MATLAB/Simulink vehicle models (EV powertrain, tyre forces, energy optimisation) with closed-loop sims and telemetry checks.

Engineering Intern - Haas F1 Team

(2024-06 - 2024-06)

• Contributed to model-based validation in MATLAB/Simulink; transient-response analysis and sensor data cleaning.
• Modelled wiring looms in Dassault 3DEXPERIENCE; ran EMI-reduction tests in LTspice, cutting sensor noise by ∼30%.

MEng - Engineering Science - University of Oxford (2022-10 - 2026-06)

A-Level - Christ the King Sixth Form College (2020-09 - 2022-06)