Experience
Professional experience and career history
Roles
I am building real-time computer vision products for detecting identifiers in complex real-world scenarios across multiple video streams and images. I develop web services, ML pipelines and infrastructure as code, implementing comprehensive observability with Grafana Cloud (Prometheus, Tempo, Loki). I have built benchmarking tools with data version control, document analysis solutions leveraging LLMs, and internal tools in TypeScript and React. I contribute to API servers in Go and have project managed teams, driving planning and development practices.
I led our ML strategy in this pre-product-market-fit, seed-stage start-up. I built ML pipelines for training and evaluating models, experimented with 3rd party services to evaluate "build vs buy", and built PoCs to discover our desired UX. I built web services in both Python and TypeScript and contributed to iOS development. I built CI/CD pipelines and infrastructure as code.
I designed, architected and engineered machine learning driven data streaming solutions, wearing multiple hats as software engineer, product owner, project manager and technical consultant.
As Sonrai's ninth employee, I was central to establishing its engineering practices. I acted as the technical lead for our Data Science team, making many of our foundational architectural decisions and training our team of junior engineers. I developed computer vision models for semantic segmentation, multiple instance learning and object localisation. I engineered full machine learning pipelines, managing the scope of algorithms, data sampling, modelling, versioning and iteration.
I led our development of production-grade computer vision solutions, including algorithm conception, data collection and engineering, deep learning model development, developing a scalable runtime environment and integration with web applications.
I worked on the SPAAACE project (Self-Properties Autonomic / Apoptotic / Autonomous Computing Environment), using Arduinos to create a prototype of Cubesats implementing university research on self-managing systems.
I primarily designed, developed and tested C# code which improved auto-suggest across all platforms. I optimised large portions of the stack, carried out in-depth big data analysis on logs, and developed interactive dashboards.
I developed a system which took the results of NLP analysis of web-pages, and generated a graph database to hold semantically related information to the web-page contents.