Daniel Ralley
Computer Science at Southampton Solent University
I'm building a career at the intersection of production software engineering and machine learning – the kind where models actually make it from Jupyter notebooks to real-world impact. Having just completed my MSc in Applied AI and Data Science (with Distinction!) while simultaneously leading Python development on a university research project, I've discovered my sweet spot: full-stack ML systems that solve meaningful problems. My path here wasn't traditional, but it's given me something valuable. Five years shipping production-ready embedded software - globally at Dyson - taught me the discipline of containerised CI/CD pipelines, 80%+ test coverage, and cross-functional collaboration under pressure. Now I'm channelling that "production DNA" into ML deployments – building healthcare platforms with FastAPI and Docker Compose, implementing deep reinforcement learning for robotics (with GPU optimisation), and developing computer vision pipelines that bridge photogrammetry with automated 3D character rigging. I love the challenge of taking unique concepts and transforming them into robust, scalable systems that people can actually use. What genuinely excites me? Projects that create positive impact. I'm drawn to climate tech, healthcare innovation, and automation that improves lives rather than just optimising metrics. My recent work includes building an AI virtual receptionist that's measurably reduced patient hold times, creating healthcare platforms that make medical information accessible across language and cultural barriers, and developing 3D computer vision tools that could revolutionise animation workflows. The common thread: technology solving real problems for real people. I'm particularly passionate about the intersection of computer vision and 3D spaces – there's a curious niche, teaching models the intricacies of posing 3D characters. Ideally, going from photorealistic ground truth to a posed 3D mesh, could show some awesome potential. If you're working on projects that combine ML with meaningful impact (especially in climate, healthcare, or 3D/spatial computing), I'd love to connect. Let's build something that matters.
Work by Daniel Ralley
- Justin - The Voice AI Receptionist
Fun story: originally this was meant to be a project to help me learn some Node.JS, TypeScript and N8N - now it's a fully-serviced product working out-of-office hours for friend's business! This is a low-latency and cost-effective virtual receptionist solution, leveraging OpenAI's real-time API…
- Pose-Space Deformation with ML: Image -> 3D Mesh -> PSD Blend Shapes
https://github.com/RalleyD/scan_to_smpl This project, introduces an automated end-to-end pipeline designed to transform subject images into near-production-ready animation blend shapes. By bridging the gap between body scanning and functional character rigging, the system utilises a hybrid…
- Walking with DRLs
https://github.com/RalleyD/walking_with_DRLs This project provides a comprehensive comparative analysis of foundational and advanced Deep Reinforcement Learning (DRL) algorithms applied to the complex challenge of simulated bipedal locomotion. By developing a rigorous A/B testing framework, the…