About

My Journey in AI and Data Science

My path into AI started during my maths degree. Lectures on linear algebra, probability, and statistics unknowingly built the foundation for a career in AI.

My first role was as a Software Engineer at IBM. Pure software wasn't my natural fit, so after three years, I switched to AI, where I've been ever since.

I'm an AI generalist, having worked across various industries and across the full AI lifecycle - from research to production.

With 15 years in tech and over a decade in data science and AI, my focus has been shipping AI solutions with real business impact. My transition from Mathematician to Software Engineer, then to AI leader, taught me what works in AI and what doesn't.

Lessons That Shape My Approach

Here are a few lesson I picked up along the way that guide how I approach my AI work:

Focus on the Right Data, Not Just More Data

At IBM, I learned quality data often beats big data. This principle still shapes my work.

Start Simple, Measure Everything

At Elsevier, working on recommender systems highlighted the power of simplicity. Simpler baselines are crucial before exploring complex models. Elsevier also taught me to link internal metrics to business results, like customer satisfaction and retention.

Bridge Tech and Product

Even the most advanced tech isn't of much use if it doesn't end up in a product that helps people. So, I've learnt to work collaborate closely with product teams. The goal is always to turn technical innovations into real customer benefits.

Respect the Data Foundation

Building transaction classification models at Experian one again drove home a key point: good AI needs good data. Proper data annotation and understanding are foundational. No amount of advanced modelling can make up for poor data.

Adapt to Constraints

At Quin, a med-tech startup, I had to adjust ML methods for on-device limits and tight regulations. That was valuable experience for learning how to adapt situations where AI has to work within tough constraints.

Lead with Purpose

Leading teams at Monolith AI - anywhere from 4 to 20 people in data science, ML engineering, and research - helped hone my leadership style. I've found that small, focused teams with clear goals and business-linked metrics tend to do better than larger, less focused groups.

My approach, put simply is: Understand your data. Build simple things first. Measure what matters to the business. And always, always focus on customer outcomes. I've found this way of working delivers AI solutions that works best in the real world.