I worked as a Product Design Intern at Baidu natural language processing Department to create a cross-platform simultaneous interpretation service. I participated all cycles of the design process from research and UX/ UI design through QA , all from scratch.
This platform is a new version in Baidu AI Simutaneous Interpretation solution, which aims to provide support for large and medium-sized international conferences that require high-quality performance. We hope to integrate Baidu's all-round technical capabilities, integrate the product framework, and gain experience and improvement from our previous product.
2 Product designers, 2 product managers and 10+ engineers.
5 months
March 2021— July 2021
UI/UX, Wireframing, Rapid Prototyping, User Testing, Strategy
With VIP Version, user can set up and enjoy real-time transcribing, translation and on-screen display in multiple languages. They can also use integrated services such as conference management system, terminology customization, artificial subtitle guarantee, and conference minutes.
When users want to handle a large and medium-sized international conference, the experience start breaking when dealing with complex environment and meeting content, since The “Ordinary Version” originally designed for small-sized businesses. Then, if users turn to the “Customized Version”, they need to pay 100 times more than the original outlay, and our team needs to occupy 2-3 staff to support the meeting.
How might we design a simultaneous interpreting system to create a seamless user experience and enables independent usage, thereby supporting medium and large international conferences?
Existing speech recognition and machine translation technologies cannot guarantee accuracy
Many functions and settings are distributed in the different interface or module.
Users need to complete the process of meeting creation, setting and execution on multiple platforms
Baidu AI Simultaneous Interpreting followed dated UI paradigms from an old tech stack, with critical usability issues.
Aside from the basic competitive research, I accompanied by customer support lead and PM, conducted 10 interviews with our in-house conference operation specialist and business organizers trying to stay unbiased and as neutral as possible. We were looking forward to checking and pointing out common pain points from the quotes ,and translate them to clear goals.
"The content of the meetings of the same company is similar, but I need to reset the screen and enter the relevant materials for each meeting, which makes me spend too much repetitive work time."---Conference operation specialist 2
"To support a high-quality conference, we need to connect multiple devices and adjust the simultaneous interpretation system repeatedly. In the absence of a clear operating guide, this is a work that requires learning costs."---Business Organizer 3
"During the meeting, I need to observe the dynamics on the stage and monitor the status of simultaneous interpretation. In case of emergencies, the scattered functional modules make it impossible for me to make adjustments in time."---Conference operation specialist 5
Enhance operability to adapt to users with different cognitions
Simplify page hierarchy and flow, helping users quickly support meetings
Focus on improving the accuracy and quality of simultaneous interpreting
Users can quickly setup meetings on the home page, eliminating the flow with high drop-off rate during setup. The function modules arranged in the home page according to the demand priority.
The console is a tool page that collects all meeting settings. We divide areas according to functional attributes to help users efficiently control the meeting.
All the control functions are arranged on the right side toolbar. Users can click the arrow to expand or collapse the toolbar.
Based on the theme of the meeting, users can enter the spelling, pronunciation, translation of the terms related to the meeting in advance. Four customization modes effectively improve the accuracy of speech recognition and translation.
We introduced general glossary and conference glossary. The general glossary is effective for all meetings, and the conference glossary is only effective for a single meeting. Partition management helps users save a lot of input time.
Users need select the customization type and language, and then enter the content. The UI layout and UX writing of each mode are fine tuned according to its goals and specific restrictions to make the input work more accurate and efficient.
We have prepared a term template. Users only need to input all terms into the form, and then they can upload terms in batch with one click.
Importing other existing conference term libraries can realize multiple conference shared glossary, helping users manage conference glossary quickly and personalized.
This is a management center that shows all meetings. Users can view the status of the meeting, the duration of simultaneous interpreting, and the minutes of the meeting.
Through the usability test, we chose the typesetting and paragraph length that users are more suitable for, and designed two reading modes of meeting minutes, "full-text comparison" and "bilingual comparison".
Mobile phones can have better radio effects than computers, so we made the app radio terminal to follow the design principles as simple as possible.
I'm so lucky to be able to participate in the whole design process of an influential product from creation to launch. This challenging internship taught me how to collaborate with diversified teams and face multiple tasks in different stages. With the experience with document organization and meeting communication, I learned to speak with engineers and follow their needs through rounds of iterations.
This project gathers the latest and most advanced technologies in simultaneous interpreting industry. I spent a lot of time learning the underlying logic of the technology we use and the technical limitations that accompany it. With the help of product managers and technical team, I created scalable designs that transferred complex technology into simple operation experience, resolved user paint points and edge cases.
For more work inquiries, or to grab a coffee do email me at
wanqiu0930@gmail.com ☕️✨.png)
2023 Wanqiu Wang. All rights reserved.