About
I'm a data scientist who got restless staying in one lane. Not out of restlessness for its own sake — but because the interesting problems rarely fit neatly into job descriptions.
I work in consulting at Publicis Sapient, embedded in feature teams building a GenAI platform for a global CPG client. In practice, that's meant fine-tuning image models on brand IP, migrating pipelines to microservices on Kubernetes, writing specs, running demos, and making the case to stakeholders for why the trade-off you're recommending is the right one.
The range across AI/ML, infrastructure, and product delivery is deliberate. AI is moving fast enough that the most valuable thing isn't picking a lane — it's building judgment. That's what I'm working on.
Credentials
Scroll to see all →
Work
Enterprise GenAI work, shipped. Full case studies in progress — what's here gives enough context to understand scope and impact.
Part of the feature team that fine-tuned image models on client brand IP — enabling AI-generated marketing visuals at scale across 15+ global markets. Owned technical delivery end-to-end, from model iteration to stakeholder demos.
Identified model-to-production latency as a bottleneck nobody had named yet. Proposed and led the shift to modular microservices — cutting deployment time by 50%+ and unblocking rollouts across markets.
Thinking
Not a blog. These are structured teardowns — products I think deserve more honest analysis than they usually get.
Contact
I'm open to conversations about interesting AI/ML problems, technical roles where scope isn't fully defined yet, and early-stage teams that value range over specialisation. If something here resonated — reach out. Happy to talk.