About Me.
I'm Mark Stent. Data scientist, music producer, and someone who's been quietly obsessed with both for long enough to stop apologising for the combination.
The day job is leading data science and AI work across multiple markets. I'm deep in the generative AI side of things: benchmarking models, prototyping with LLMs, building agent workflows, and working out where RAG, synthetic data, and predictive modelling actually do useful work versus where they just look clever on a slide. I lead and mentor teams, shape AI strategy, and spend a fair bit of time helping non-technical people get properly capable with these tools instead of just impressed by them. Before moving into leadership I taught machine learning, GenAI, and applied data science to several hundred learners across university and apprenticeship programmes, which is where I learned to explain hard ideas without dressing them up.
On paper: MSc in Computer Science and Data Science from the University of Sunderland (Distinction, with a dissertation on audio diffusion and generative music, yes, really), a BSc in Mathematics and Statistics, and a Diploma in Datametrics and Statistics from the University of South Africa. Stanford's Generative AI programme in 2025. Member of the Royal Statistical Society, with a paper in preparation on bonus optimisation in gaming.
The music came first, and it's not going anywhere. I've been DJing and producing since 2005, mostly in House, Tech House, and Techno. "Come Back Home" sat at number 1 on 5fm for 12 weeks and finished as the overall number 1 on the 2016 Top 100. Solitude picked up a SAMA nomination for Best Pop Album in 2020. I've headlined Ultra South Africa, H2O, and Aquafest, shared stages with Armin van Buuren, Fatboy Slim, Swedish House Mafia, Tiësto, and a multitude of others, and played through Bahrain, London, Ibiza, Portugal, and Dubai. I run three independent labels (Stentfire Music, Flexual Records, and 3 Girls Music) and have released techno through BPA, the European agency behind Richie Hawtin and John Digweed.
The two halves aren't as separate as they sound. Both are experimentation work. Both start with a hunch, a rough version, something that's nearly there, and the willingness to bin it and start again. Releases and code are linked below. Always up for a conversation about either.
