Role
Product designer
Timeline
August, 2022 - Present
Group
Caoyang Shen
Yin Li (Founder)
Tools
Figma
The case study below may be an old version because we keep updating as a startup. You can go to the website to see more.
Background
As AI becomes more specialized, AI projects often fail (80% of the time) due to a skills gap.
No quality, cost-effective solution for fast and flexible hiring.
A complete career path and flexibility do not coexist for AI talent.
The current recession makes the company have to reduce its budget. The freelancers will be good helpers for them to try the new projects at lower costs.
Existing flexible work solutions fall short
Design question
How might we help clients find the best freelancer team for their AI projects?
How might we enable engineers to work more flexibly within a stable team?
See what SylphAI does to solve the problem
Quick Look
Personas
For clients
The platform should allow clients to enjoy the quick and accurate matching of candidates for their projects.
Clients can search teams, learn about and select the candidates and finish the contract in a one-stop line.
Direct talking and cooperating with engineers can save time and budget.
Demand: Easy and accurate to find the right team.
For junior engineers
Junior engineers can conduct more projects with the help of senior engineers.
The team makes the networking not stop at talking.
Demand: Gain experience as an AI engineer in a stable team
For senior engineers
Demand: Have a chance to do more challenging and interesting thing
We want to decentralize the industry, making engineers create their ideal team without relying on consulting firms.
The engineers can choose their favorite teammates, making the work efficient.
Meanwhile, engineers can have more initiative to select their suitable projects.
There are three key players on our platform. Clients post projects, teams apply for projects and team leaders build teams.
Market research
01. Clear Projects lead to accurate matching.
Unlike competitors, we request clients have a clear idea about the project. It enables engineers to know whether they are suitable. The AI industry has different and professional fields, like computer vision, machine learning, voice process, etc. Traditional platforms are hard to find the right person.
02. Team increases the working stability.
Most platforms only allow clients to hire one freelancer instead of one team. Freelancers only finish the task without more consideration on the product development. However, the team model will make the freelancers access to more support. We believe that teamwork can reduce the frequency of personnel changes in the project to some extents.
Design starts
⋆
Design starts ⋆
In order to make the design process understandable, we will focus on the angle of the team leader.
Let’s meet our friend, Mario!
User flow
It is a double-sided matching process. Mario needs to find Luigi to create a team, and he needs to find a suitable Mission that is not too easy or too difficult for him.
Solutions
Team dashboard · Identity
The trip for each engineer should start from our team dashboard. We encourage them to build a team, and we believe the team will increase their identity, which increases the chance they stay on our platform. Meanwhile, the goal for each team is also clear——matching a suitable project.
Sourcing · Detail
We want to allow team leaders to browse more projects on one page. According to our interview and discussion with AI experts, we decided on the information presented in each project list.
Matching· Efficiency
Comparing projects is a time-consuming process. Our platform gives more freedom for users to sort the projects. When regret, you even can retrieve the rejected one from our rejected area.
Teammates selection· Flexibility
We allow team leaders to recommend their teammates to clients. During the proposal, they can still return to change the teammates in this project. Each project does not require everyone on the team to participate in that.
Proposal process· Transparency
Our system helps users track each step and record the files during the negotiation. Users on each side can see the other’s process, ensuring they are on the same path.
Design Iteration
After launch, we underwent a design iteration according to the feedback of users and the founder.
We make this change based on three reasons.
Simplify users’ management of their teams and projects.
Enhance user experience, avoiding too empty a page.
Increase the interaction through red pots.
The founder's vision is to provide a transparent and flexible platform for clients and engineers.
In the beginning, we put the teammates’ selection as our final step of the proposal. The selection right is on team leaders. We used a pop-up to realize it.
Considering the legal issues, we made several changes.
Transfer the right of selection from leaders to clients. The team leaders can recommend their teammates, but the clients will make decisions.
Advance the step of teammates’ selection. In this way, teammates selected can also participate in the following proposal process.
Show all teammates to the clients. It may increase clients’ understanding of each candidate. It may help clients’ future recruitment.
Users indicated that they feel it is hard to understand the proposal process because projects and files are on the same page. We made several changes to help users track the process.
Status shows the negotiation step where two sides are.
Divide the file management into two categories: my part and the other side. Users on both sides can learn about the proposal's progress quickly.
Projects and files are divided. Users can go from the project list to manage the files in a particular project.
01. Reducing the click frequency and enriching the information on one page.
02. Reorganize the team selection, bringing more flexibility and transparency to the clients.
03. Improve the useability, making users to manage their files and understand the process easily.
Hi-fi Prototyping
Achievement
Design System
Next Step
01. Usability Improvement
During the project matching, there is confusion about teammates' selection and proposal process. Regarding teammate selection, we want clients to receive recommendations from the team leaders. However, the selection will show simultaneously to the clients and leaders, which means that clients sometimes have to make their own decisions without suggestions from the leaders. Meanwhile, we allow clients to communicate with the teams only after the match. However, it does not make sense, so we should enable them to communicate as soon as possible.
Regarding the proposal process, we only allow team members to look at the process without any uploading, considering of development difficulty. However, in the future, we should introduce AI to sort and organize the files, enabling different roles to participate in the discussion before the project starts.
Reflection
01. Learning is the first step of design
At the very beginning of the company, the designers very easy to follow their own ideas and the founder’s thoughts. Before launch, we were not able to be exposed to actual users’ data. Considering the case, the comparators analysis is critical for designers. We understand others’ designs by visiting Dribbble and Medium, then apply them to our platform. When I review my design, I always ask myself questions:
what do others do in this step?
Have I seen a similar design or user flow in others’ work?
02. User is a key but not everything
We always want to improve the user experience, especially during school projects. However, in practice, a designer will think more. Whether the developer team can understand and realize the demand, whether the benefit overweighs the cost, and whether the design improving the user experience will harm the company’s benefits. What can be conceded and what cannot is also an issue for a designer to think about.
