Machine learning solutions in manufacturing environments often involve complex workflows across data inputs, model configuration, monitoring, and operational deployment. These tools are frequently built for technical users, making adoption difficult for broader business and operational teams.

UX Designer responsible for workflow design, interface concepts, prototyping, and usability improvements in collaboration with product, engineering, and technical stakeholders.

Originally built as a Python command-line tool by a data science team in Japan, the system successfully detected rare production defects using a low-cost camera setup. My role was to transform the workflow into a structured, usable product that could scale beyond the original expert team.

I designed interfaces that simplified machine learning operations into clearer, more manageable workflows across multiple platforms.

• Guided ML Workflows
  Designed clear interfaces for image preprocessing, model training, results review, and anomaly detection.
• Faster Operational Decisions
  Surfaced critical metrics and alerts clearly, helping users assess model readiness and respond quickly in production settings.
• Cross-Market Scalability
  Customised the platform for a Korean Bosch business unit and designed an offline Python app for local deployment needs.

Many AI tools fail not because the models are weak, but because the surrounding workflows are too complex for everyday users.
By improving usability and clarity, the product could be used by a wider group of stakeholders beyond technical specialists.

Expanded from a single expert-built solution into a reusable platform adopted across multiple Bosch teams.
Demonstrated how UX can turn technical AI tools into scalable operational products.

• UX for machine learning products
• Workflow simplification in technical environments
• Enterprise multi-platform design
• Designing for expert and non-expert users
• Operational adoption of AI systems

The goal was to make ML tools more usable, understandable, and operationally practical.