CTO & Technical Lead at Bravyn - building AI-powered leadership coaching. Hands-on tech leader specialising in conversational AI, multi-agent orchestration, RAG architectures, and context engineering.

I build things end-to-end: from infrastructure (AWS/Terraform/Kubernetes) through backend systems to AI pipelines and user experience. Career spanning fintech, video delivery (local and broadcast IP), e-commerce, and now AI/ML.

Bravyn - AI leadership coaching platform with full-screen avatar-based coaching on iOS and Android. React Native, on-device speech recognition and synthesis (native modules), Python/FastAPI, multi-agent orchestration, streaming inference, LLM observability, Ruby on Rails, AWS

_{Real-time avatar coaching: on-device speech recognition and voice synthesis drive a full-screen avatar, with the live transcript over the scene and recognised speech surfaced in the voice bar.}

RefTracker - computer vision + Gemini pipeline that tracks rugby referees and explains their decisions. YOLOv8, BoTSORT/CLIP-ReID, OpenCV, Supervision

_{YOLOv8 person detection picks the referee by their unique kit (shirt) colour every frame, falling back to multi-object tracking (BoTSORT + CLIP-ReID) when the colour isn't separable. Velocity clamping to real running speeds, camera-motion compensation, and teleport detection (rejecting physically impossible between-frame jumps) keep the marker locked to the right person — who is only ~40px tall in the original 1920×1080 footage. The three panels are a single tracking pass shown at three zoom levels.}

• Effective Conversational Closure in Supportive Dialogue (with Ruyi Wang et al.) - pre-publication, targeting EMNLP 2026 Industry Track. A four-dimensional latent-space classifier that steers a coaching model at inference time, addressing maintenance bias and termination inertia. Production system outperforms baseline models across 85% of real sessions.

• AI/ML - LLM application development, multi-agent orchestration, RAG, context engineering, evaluation and safety pipelines, conversational AI, computer vision
• Infrastructure - AWS, Terraform, Docker, Kubernetes, IP networking, video delivery
• Product - 0-to-1 product architecture, UX-driven development, full-stack delivery