At the beginning of the project, I approached sales team and data scientists to conduct individual interviews and facilitate brainstorming session in group.

What I was trying to understand is:
- What business problem are we trying to help with?
- Who could be our end-users?
- Why do we need to build the product?

Based on all results, I created a persona to address all the problems.

It was my first data-centric product design. Before I got started, I put myself into a data person's shoes. I watched videos, read articles and asked my team members to understand a data analyst's work flow and get basic knowledge about machine learning and AI.

In the meantime, I also looked at other similar products and compare them regarding all available features.

After understanding all the client needs and business requirements, we defined the product scope focusing on building model monitoring tool. Also, we defined three main goals for the tool - user control, easy to observe and find issues.

I also facilitated brainstorming session with data scientists to map out user flows for the model monitoring tool including both new user scenario and return user scenario.

After defining product requirements and user flows, I started ideating the design for each page. I created variations and presented to my team members.

I kept collecting feedback from my team members and then moved from medium fidelity to high fidelity.

User can connect different models under one project, followed by specific model monitoring data under each model.

Real-time alerts are sent to users once there is an issue with the highlighted feature. Users are able to go to notification center to manage all messages. Clicking each message directs them to the specific model.