PhD (AI/ML) | MD | MBA
AI/ML engineer, researcher and consultant.
Serial founder and author.
About
I conduct research on AI systems. My work includes mammographic imaging AI with publications in Nature and IDBM, computer vision methods, and procedural memory architectures for language models. I am currently researching Depth-Elastic Recursive Transformers, which provide a framework for more efficient LLM architectures.
I have founded Alixir, Fleet, and Enagram, building products from concept through production. I have also served as founding and fractional CTO across health-tech, ed-tech, gaming, events, entertainment, and fintech, working on architecture, infrastructure, and product development.
Expertise
AI Research & Development
Conducting research on novel architectures, optimization approaches, and AI systems. I work on methodology development, model design, and advancing practice in production environments.
Production AI Systems
Building and deploying AI models and systems in production. From model serving and retrieval to the infrastructure and data layers beneath them. Engineered for correctness, safety, and trust.
AI Strategy & Consulting
Advising organizations on AI implementation, architecture decisions, and technical direction. I help teams navigate technology choices and long-term strategy with engineering rigor and real-world constraints.
Engineering & Architecture
Building systems and infrastructure for production AI. Designing distributed systems, data platforms, and technical architecture that are observable, maintainable, and resilient.
Recent Work
I have founded several AI companies and collaborated on research and applied projects across computer vision, ML infrastructure, and systems work.
Founder / Founding CTO / CTO
Fleet
Founder and architect of Fleet, a developer platform for training and deploying AI models on cloud GPUs. Developers point local Python at a model and dataset, and Fleet handles GPU provisioning, containers, storage, and serving.
Learn moreEnagram
Founder of Enagram, a memory layer for LLM applications. Built on the Continuum Memory architecture, it benchmarks against Mem0 and Zep using purely heuristic retrieval, with no LLM in the loop.
Learn moreAlixir
Co-Founder and CTO of Alixir, building applied AI systems for medical imaging in emerging economies, delivering state-of-the-art inference on benchmark datasets.
Learn moreExceptra
Founder and architect of Exceptra, an AI-native logging platform that parses and triages logs with LLMs and surfaces incidents directly in Slack.
Learn moreRaceway
Lead maintainer for Raceway, an open-source distributed systems debugging platform written in Rust that detects subtle thread concurrency issues through vector clock analysis.
Learn moreArcadian
Founder of Arcadian, an AI-native web platform for creating games and playable ads, supporting Canvas and Three.js alongside an internally built game engine.
Learn moreConsulting / Engineering
Shark Trust
Built custom computer vision models for shark identification for a UK-based shark conservation charity, turning field imagery into reliable species recognition.
Learn moreWestfield
Re-engineered and architected Westfield's e-commerce data pipeline to facilitate integration into third-party platforms without compromising compliance or security.
Learn moreUTS
Collaborated with the University of Technology, Sydney on research spanning computer vision, applied AI, and data science.
Learn moreFrom peer-reviewed research in AI and machine learning to a book challenging how we build startups in regulated industries.
Why startups fail in deep tech and regulated industries, and what actually works. A framework called the Evidence Based Venture Pathway, the antithesis to lean startup for high-stakes ventures.
Contact
Whether you need research guidance, engineering help, or technical strategy consultation, I'm interested in working on problems that matter. Tell me what you're tackling.