Languages: Python, C++, C#, JavaScript, TypeScript, SQL, PHP

ML / AI Tools: Machine learning, deep learning, PyTorch, TensorFlow, Keras, OpenCV, linear algebra, segmentation, object detection, transfer learning, Bayesian deep learning, neural networks, computer vision, image processing

Tools: Git, CLI scripting, Python packaging, Jenkins, Github Actions, PyTest, Pandas, Django, Flask, Docker, LaTeX, Linux, macOS, systemd, AWS, PySpark, Hive, Jupyter Lab, DataBricks

2017 – 2020: Bsc Physics (2:1)
Queen Mary University of London, London, UK

2016 – 2017: Access to Science Diploma (Distinction)
Tower Hamlets College, London, UK

2003 – 2007: Bachelor of Fine Arts (Photography)
Victorian College of the Arts, Melbourne, Australia

Dimension Studio, London UK, Nov 2023 - Curr.

Software Engineer - Machine Learning and Research, (Applied Technology Team)

In my role, I am responsible for researching and implementing both novel and existing machine learning techniques for virtual production and live events. This involves developing and maintaining software specifically designed for moving-image based purposes. I collaborate closely with cross-functional teams to ensure the successful delivery of projects.

Live Production - Omega, Race of Champions (Omega Pavilion, Parc de Bercy, Paris Olympics, 2024): Brand campaign featuring live audience interaction with ML segmentation and matting from multi-angle video capture, racing stats overlay, and composite next to a 3D virtual athlete (meta human), rendered in near real-time.

• I Developed a pipeline with two simultaneous Linux worker nodes:
  • ML segmentation/matting of Human participant from multi-camera feeds
  • Video composite of rotoscoped participant onto self selected pre-render 3d virtual athlete (meta human), with race stats overlay
  • Near real-time processing: tasks were added to a queue and delivered within 5-10 minutes of audience participation
  • Software stack: Python, OpenCV, PyTorch, Linux, systemd
• Featured on Omega’s social media and website
• Audience count: 500+ per day, for duration of Olympics and Paralympics 2024
• Athletes: Noah Lyles, Shericka Jackson, Hannah Conckcroft, Marcel Hug

Expedia, London UK, Feb 2022 – Nov 2023.

Machine Learning Engineer II (Meta Marketing Technology Team)

Cross-functional team supporting and maintaining customer prediction models and software platform running on AWS

• The platform is built using Airflow, Hive, Python and runs on AWS EMR instances, with Jenkins pipelines for model training and deployment, and DataBricks
• Platform triage support for users located in many time-zones
• Lead MLE in project migrating legacy project w/ linear process into AirFlow DAGs, enabling step tables, concurrent processing, and code quality improvements

Arm, Cambridge UK, November 2020 – Jan 2022.

Software Engineer, Machine Learning Group (Graduate Software Engineer 2020 - 2021, Software Engineer 2021 - 2022)

Taking part in the graduate rotation programme exposed me to a variety of different projects and teams within the machine learning group at arm:

• ML Research team, Bayesian DeepLearning for CV: (2 months) Lead researcher, exploring and implementing Bayesian DeepLearning models for pixel level image segmentation, and optimising for Arm hardware:
  • Explore model architectural changes for performance optimisations
  • Based on DeepLabV3 architecture
  • Deep dive into Bayesian DeepLearning methods with a model implementations for pixel segmentation.
  • Model implemented using Keras and TensorFlow
• Applied ML team, Dynamic hand gesture recognition, and visual wake words (8 months): An ML model for static and dynamic gesture recognition, aimed at low power / IOT / embedded processors
  • Lead researcher exploring SOA techniques for dynamic gesture inputs
  • Research and read papers on novel CV ML model architectures and datasets
  • Create data generator pipeline using OpenCV for large video dataset
  • Implement novel architecture aimed at low power processors
  • Report metrics for comparing model accuracy, peak memory usage, FLOPS on base model to compare on a variety of arm hardware IP, and showed changes after model quantisation and other optimisations
• ML Tooling Team, IPSS-ML (IP Selection Sandbox for ML applications) (4 months): a middle-ware application to simulate and test ML applications on accelerated (NPU) and regular (M/A class processors) Arm IP using fast- and cycle-models.
  • Increase test coverage to all lines of code, adding niche test cases
  • Bring codebase to be completely error and warning free by implement Pylint into CI environment
  • Bug fix front end web app
  • Deliver presentation on methods used to implement Pylint into a live codebase, and Jenkins CI build process with Gerrit

Deimos Space UK, Harwell, Oxfordshire, July – Sept 2019. (Internship)

Computer Vision and Machine Learning research intern

• Built object detection model using neural networks for earth observation data, to detect and differentiate between biodiversity types (Python, TensorFlow and Keras)
• Researched methods of transfer learning in neural networks, and implemented one in a project constrained by a limited labelled dataset; improving training time
• Created internal reference documentation for CVAT (Computer Vision Annotation Tool)
• Collaborated on a poster researching computer vision methods using machine learning, neural networks and transfer learning

Yobota, London, Jun – Sep 2018. (Internship)

Software Engineer intern

• Developed integration API for OpenBanking using Django / DRF / OAuth 2.0
• Implemented unit tests using pytest and integrated with continuous integration tool (Jenkins)s
• Delivered team keynote on OpenBanking
• Create documentation on “Creating a test driven integration” to work on the Yobota platform
• Participated in daily stand-up and weekly sprint planning meetings

(Selected freelance)

• June 2016 – 2020: Stillnessinyoga, Remote, The Netherlands
  Web developer / Digital content development
  • Lead developer for multi-lingual e-commerce WordPress platform
  • Built and maintained website using the Genesis theme framework featuring an events calendar, e-commerce platform, content restricted membership subscriptions, and membership e-learning course and system
  • developed a custom secure Amazon S3 media serving platform
  • Worked along side head teachers to design and put together three teacher training manuals (Level 1, 2, & Advanced 2) – in iBook format – an interactive e-publishing book format
• 2016: Ecostage Pledge. Web developer. Designed and coded the website for a community of artists and designers. Features include a membership which integrating BBPress user interaction features
• 2015: Giddy Diva. WordPress theme developer. Migrated a static site to WordPress, Built out theme, integrated client testimonials.
• 2013 – 2015: LVSC. Web developer. Several projects including VCS Assist, London For All. Designed and coded the community website for publishing a blog and news articles. Later created a business directory, for users to publicise their projects
• 2012: London for All. Web developer. Designed and coded the community news site
• 2012: MIAGOA. WordPress theme developer. Developed site from a photoshop image design template.