K1st BUILD
Building AI models from human expertise
Problem
Manufacturing companies struggled to build AI models due to incomplete datasets, requiring significant time for data preparation and delays in operations.
Solution
We created K1st BUILD, a knowledge-first AI model building app that leverages expert insights, reducing reliance on datasets while accelerating model development.
Reduced data prep time by 30%
Improved model accuracy by 15%
Company
Aitomatic
Position
Founding Product Designer
Year
2022 – 2023
Market
APAC & US, B2B
Problem
Manufacturing companies often struggle to build AI models because their datasets are messy, incomplete, and unreliable. AI engineers spend weeks piecing together scattered data just to prepare it for use. This slow, error-prone process delays model development and risks costly downtime.
The key challenge was building AI models fast and accurately when the data itself wasn’t perfect.
My role
As the Founding Product Designer at Aitomatic, I was responsible for transforming expert insights into a structured workflow that could compensate for the shortcomings of imperfect data.
Working closely with CNA, a Panasonic Group company, and our engineering teams, I understood challenges faced by AI engineers and domain experts. I refined the user experience through rigorous testing, ensuring that K1st BUILD could translated expert knowledge into AI models.
What we built
I designed K1st BUILD as a knowledge-first platform that integrate domain expertise into AI models.
The platform was built with two distinct user personas in mind. The primary persona is AI engineers, who build the models and need a streamlined process that reduces data preparation time. The secondary persona is experienced domain experts, aka industrial engineers with 30 to 40 years of operational insight. They can contribute critical knowledge to enhance model performance.
Instead of relying solely on raw data, both groups can document and integrate their insights into structured, reusable knowledge that fill in data gaps. K1st BUILD supports data schema and real-time refinements of models on the fly, ensuring precision needed.
Through testing sessions, I fine-tuned the workflow to ensure that expert inputs flowed effortlessly into the AI model development process, creating continuous feedback loops between AI engineers and domain experts that drive ongoing validation and improvement.
Impact
K1st BUILD transformed how manufacturing teams approach AI model development.
By combining human expertise with data, K1st reduced data preparation time by 30%, enabling AI engineers to focus on building models rather than pre-processing data. In use cases like predictive maintenance, model accuracy improved by 10-15%, delivering results in weeks instead of months.
Lastly, teams reported about a 20% improvement in alignment when sharing data, as K1st BUILD facilitated smoother collaboration between AI engineers and experienced domain experts.
