Automating Operations with Machine Learning

From weeks to days: making machine learning accessible without code

Services

Product design · UX research · Roadmap & success metrics definition · Cross-functional collaboration (engineering & data science)

Description

Built a platform that simplifies how ML models are created, run, and monitored. The goal was to help both technical and non-technical teams work with machine learning systems without requiring deep ML expertise.

Impact

Weeks → days reduced onboarding time
4 new clients deployed
Made complex ML workflows usable by non-technical business users
Defined product direction and roadmap alongside the tech lead

Year

2024

Timeline

Growth

Location

London, UK

PROBLEM SPACE

The delivery gaps this work had to close

Managed Analytics was too technical for the business users it needed to serve, and the team building it was missing key support.

Technical Barrier & Current State

Some clients lacked the technical expertise or data science teams needed to build, manage, and sustain AI solutions with the existing tools. The initial version of MA had a high technical barrier, needing a redesign to become genuinely intuitive and user-friendly.

Team Dynamics

Two months after I joined, a leadership gap emerged with no PM in place, alongside limited understanding of the user. The frontend team was also struggling to implement designs, which gave me the opening to step in on both fronts.

Not Built for Purpose

Users found technical terms like 'config' and 'parameters' confusing, and felt overwhelmed by the volume of information on screen at once. Certain interactions, like navigating the data catalog, created friction, and many users preferred engaging directly with output data over abstract settings.

Left, what MA looked like originally, right, some research synthesis/audit on the journey

CONTRIBUTION

Where I helped turn complexity into momentum

As design lead, I ran the research, shaped the roadmap, and stepped into engineering and PM gaps to keep the work moving.

Experience Design and Reducing Cognitive Load

Simplified terminology and navigation directly in response to research, cutting the friction users had flagged.

Product strategy

Defined product direction, roadmap, and success metrics alongside the tech lead. Established regular user interviews and playbacks, building a research-driven approach across the team.

Cross-functional delivery

Worked closely with engineers and data scientists to rapidly scope and ship an MVP, organising check-ins to bridge gaps with frontend engineering. Managed sprints and daily stand-ups alongside the Tech Lead in the absence of a PM, keeping delivery on track through the leadership gap.