About Me
I’m a Senior MLOps Engineer at Fuzzy Labs.
Lead/Senior MLOps Engineer - Fuzzy Labs - Sept 2022-Present
Responsibilities
- Lead the end-to-end delivery of machine learning projects from architecture design to implementation in diverse domains, including national security, education, and e-commerce, as well as internal tooling.
- Mentored and developed engineering talent, actively enabling opportunties for team members to progress in their careers.
- Partnered with the commerical team to identify machine learning application opportunities for prospective clients and translating them into tailored proposals.
- Spearheaded content creation, including planning, co-/authoring technical articles, and translating complex machine learning and engineering concepts into engaging, publishable blogs.
Notable Projects
- National security research projects on LLM safety and security, focusing on model provenance and attack mitigation. Involved working closely with research scientists to build out ideas. Technical work included building infrastructure and RAG-based systems from the ground-up. Tools: Transformers, vLLM, ZenML, Vector Databases, Experiment Tracking, Kubernetes, and Azure.
- A RAG-based mental health conversational system, involving writing data pipelines to process data, prompt engineering, model evaluation and monitoring, and implementing guardrails, with a user-friendly UI. Tools: ZenML, Transformers, Guardrails.ai, MLflow, Kubernetes, and Azure.
- Working in close partnership with a UK police force to investigate the applicability of generative AI in aiding officers on the ground. Included building proof-of-concepts with Azure-based Speech-to-Text services and using Large Language Models for summarisation.
Other projects have included creating open source MLOps infrastructure stacks (e.g., https://mymatcha.ai/) for traditional and generative ML, RAG-based solutions in the education domain, and building an NFT recommendation system.
Senior MLOps Engineer
Sept 2022 - Present
Postdoctoral Research Associate
Lead a project with a start-up to build data pipelines to enable content recommendation based on user behaviour with a mobile app. The app is a work-based e-learning tool used by employees to learn skills relevant to their roles within the company.
Feb - Sept 2022
PhD
The theme of my PhD is investigating how user interactions (such as clicks, keystrokes, and touch gestures) can be modelled to garner an understanding of the user’s experience with online and broadcast media. Currently, I’m exploring how large logs of interaction data, collected from live production experiences, can be used to discover predictive features of engagement and abandonment.
The application of my work is to:
- Have a positive impact on the user’s experience while consuming BBC content;
- Inform content creators through data-driven recommendations;
- Provide the BBC with an in-depth picture of the user’s experience through understandable metrics.
As part of my PhD, I’m lucky enough to be embedded into a development team in BBC R&D, where I work closely to shape how and what data is collected from live systems.
I was supervised by Dr Andy Brown (BBC R&D), Dr Caroline Jay (University of Manchester), and Professor John Keane (University of Manchester).
Sept 2017 - January 2022
Pre-PhD
Before starting my PhD, I achieved a first-class MComp (Hons) degree in Computer Science at Newcastle University specialising in Big Data & Cloud Computing (2017).
While at Newcastle, I was a Research Assistant (Summer 2017) in the Scalable Computing Group on efficient entomological monitoring of dengue fever, under the supervision of Dr Paolo Missier, where I worked on evaluating a user ranking algorithm that was developed as part of my master’s and bachelor’s thesis. The position formed the continuation of an academic collaboration between Newcastle University and PUC-RIO.
For more detailed information, please refer to my CV.