CrossClassify
Machine learning tool to protect business against fraudsters
July 2022 – August 2022
Overview
CrossClassify uses machine learning and deep learning to detect any suspicious identity or activity in any online business from the first interaction and help to makes accurate and real-time decisions.
My Role
As an experienced professional, I should help the team get the first release out as quickly as possible to demonstrate the project's capabilities. Visual design elements such as illustrations and icons were to be left for the second phase to save time. My tasks were:
-
Conduct research
-
Approve the concept with the client
-
Develop the design system
-
Design and assemble the main screens
-
Teach a junior designer the basics of working on a large project
Research & Wireframes
First, I had a good reference point — the product manager's wireframes:
After studying them in detail, I registered for similar services and studied them from the inside out.
The next step I took was to break down the user flow of a similar service to better understand the logic of the service that was already working. This also provided a good reference for the junior designer:
During the research process, there were questions and suggestions that were discussed with the client. After approving my ideas, I made some changes to the user flow and made other minor improvements.
Color Scheme & Structure
Once the wireframes were approved and it was clear what elements would be used in the system, I created about 10 mockups with different colors and element positioning:
We chose a light theme, a dark one was planned to be added later:
The final color scheme consists of shades of blue with a predominance of white, which is ideal for working with large amounts of textual data.
Design System
After approval of the color scheme and structure, I moved on to creating the design system. Each element was developed in every possible state:
User Interface
Each screen was assembled with great attention to detail:
I suggested that some of the screens be made by a junior designer, while explaining the principles of working in Figma. I recording the screen with my comments + describing the tasks in text:
As a result, all the main screens were assembled, and the junior designer gained valuable experience working on small parts of the interface under my guidance.
We discussed design details with the product manager and went through several iterations.
Outcomes
-
I mastered the subject of machine learning
-
The client got a modern design system that can handle 90% of the tasks in the future
-
The junior designer became more experienced after receiving about 20 useful video lessons from me
-
It was a great pleasure for me to share my experience
-
The developers received designs within compressed timelines and started working on the beta version of the product