Currently, I lead development of Hyperion Service, Canva's next-generation KV database (analogous to DynamoDB at Amazon) that all services will migrate to depend on in the near future. Built with a focus on multi-region migrations and BYOK capabilities, Hyperion enables Canva to serve EU customers with data sovereignty compliance, scaling our MAU significantly and closing enterprise deals with companies like Amazon where BYOK is table stakes.

I also own and operate Media Service, Canva's central media processing system that handles hundreds of thousands of requests per second and manages petabytes of data, plus lead development of the Annotation Service, which powers all of Canva's AI features and research by managing our machine learning data at massive scale with sub millisecond performance.

Previously, I launched Canva's Public REST API and implemented our in-house OAuth 2.1 system. A highlight project was CanvaGPT -- a ChatGPT app announced at OpenAI DevDay by Sam Altman. It became the #1 most used GPT at launch, serving millions of users and driving significant user acquisition. I was selected to represent Canva at GitHub Universe in San Francisco and spent time visiting the OpenAI office to collaborate on future work.

• Leading development of Hyperion Service, scaling to hundreds of thousands of QPS with tier 0 on-call responsibilities due to critical downstream service dependencies
• Architected multi-region data migrations and BYOK infrastructure enabling EU market expansion and enterprise deals
• Scaled Media Service to handle hundreds of thousands of QPS across global infrastructure
• Personally developed Canva features that closed 10 of the largest enterprise customers to date and led to direct revenue growth in the millions of dollars
• Lead development of and own Annotation Service supporting all of Canva's AI functionality at scale
• Implemented Canva's OAuth Service backing all enterprise functionality

Founded and developing an LMS-integrated AI platform positioned as "The Operating System for College in the AI Era" by addressing critical market need where universities lack visibility into student AI usage. Currently building proof-of-vision partnerships with University of Queensland and University of Sydney as part of Y Combinator application process. Backed by UQ Ventures and gaining sales traction.

• Architecting course-aware AI system using @reference methodology for lecture videos, assignment rubrics, and academic resources
• Building real-time analytics platform tracking student interaction patterns and learning behaviors for academic integrity monitoring
• Developing LMS integration layer supporting Canvas, Blackboard, and other university systems
• Implementing university SSO authentication with white-labeled interface maintaining institutional branding
• Creating usage audit logs and risk assessment algorithms to distinguish constructive AI assistance from academic fraud
• Designing scalable system targeting $2B+ addressable market across 4,000+ degree-granting institutions in the US
• Validating product-market fit through extensive faculty and teaching assistant interviews across multiple universities

My work at the University of Queensland focused on leading design and development of the first ever robotics system to perform automated Craniofacial Superimpositioning – the process of identifying missing persons by superimposing a photo of the victim with a given skull. Work was done in C++ and Python, of which culminated in me becoming a published author in the International Journal of Legal Forensic Medicine. These days, some of my spare time is taken up working on my next article on the topic.