My CV

Geoffrey Sheir

I’m a versatile software and AI consultant, who delivers tangible results in high pressure environments across elite sport, healthcare and the wider consulting industry.

Data Science Consultant, Arsenal Football Club

2026 - present

• Dedicated focus on our Women's First Team, working across match and opposition analysis, recruitment and player development.

Software and AI Consultant, PA Consulting

2021 – 2025

• Increased the precision of a smart beard trimmer by 100x using Edge AI.
  • Deployed a PyTorch convolutional neural network on an STM32 microcontroller using C, achieving sub-50ms latency from inference to motor control.
  • Reduced the latency variance from +/- 20ms to +/- 0.25ms by conducting detailed profiling and optimising the context window and memory layout of the model.
  • Worked with electronic, mechanical and optics engineers to build the prototype, as well as mentoring a software teammate as software lead.
• Built a predictive operating theatre scheduling and patient tracking dashboard for the NHS which increased the number of patients treated by 1200 in 3 months.
  • Designed and implemented a performance monitoring module for our ML model to facilitate long term support. Defined performance metrics which made the model interpretable to clinical staff.
  • Built data pipelines in Python, Databricks and Azure to combine patient data from hospital systems, and created a common data model for all Trusts to increase scalability.
  • Led the technical development of a core module, working with product managers and users to define features, and worked closely with NHS users and technical peers to test the tool.
• Deployed a GenAI PoC to automate assessment of environmental impact reports against GOV.UK policy.
  • Implemented a structured scoring system and used OpenAI foundational models via API to score reports, using quotes from the report as justification.
  • Extracted report metadata and text contents using Apache Tika.
• Upskilled 40+ NATO operatives in data and AI through classroom training and workshops.
• Pitched and won £50k of investment in a hackathon with the British Army to develop a next-gen casualty management app. Deployed the app on a training exercise and received positive feedback from Army users.

1st Team Performance Analyst, Richmond Hockey Club

2025 - present

• Established the club’s first performance analysis workflow, collecting and analysing video and event data for in-game and post-match analysis in the 2nd tier of England Hockey League.
• Presented weekly match report dashboards and video presentations for head coaches and players.

Imperial College London: MEng Aeronautical Engineering (1st Class Hon.)

2017 – 2021

• Thesis: Developed deep learning models to classify for arrhythmia-inducing drugs in pig and human hearts
  • Trained and evaluated PyTorch models using the ResNet architecture, achieving 96% accuracy.
  • Optimised the performance of the model by adjusting hyperparameter and regularisation settings.
  • Used class activation maps to identify features that the model picked up on for interpretabilty.

Personal Projects

• Assessed AWFC’s corner routines in 2025-26 using a custom data-driven approach: Read here →
• Analysed AWFC’s pressing shape in the 2024-25 UWCL Final: Read here →
• Created visualisations using Statsbomb Open WSL data: View on GitHub →

Women’s Club Coach, Cambridge City Hockey Club

2023 – 2025

• Coached our Women’s 6th and 7th teams to top 3 finishes in their respective leagues, working with players aged 14-70, delivering weekly training sessions and coaching games from the sideline on a matchday.
• Guided players through goal setting and creating individual development plans to improve their hockey.