“Our aim is to build a smarter and more inclusive world with AI.”

Gong He (MSBA ’23) is a technologist whose Deaf friends kept surfacing the same problem: text-based communication only goes so far. It carries a message, but not always the meaning, and that gap stayed with him.

Realising the limits of text as a stand in for conversation, Gong He started looking for a way to put his technical background to more direct use, one that could close that gap rather than work around it.

“We started to think about whether we can use our tech background to do something different. That’s the beginning of the whole journey,” he recalls.

He brought in four co-founders, Zeng Zhen, Xu Xuanqi, Xu Xuanzhuo and Sun Hao, and together they set out to build a way to bridge the communication gap between Deaf and hearing communities. FingerDance was the result.

FingerDance’s technology centres on a proprietary Sign Language large language model, FingerDance Sign LLM, paired with an AI avatar system. Built on generative AI trained on a diverse collection of real world sign language data, it translates spoken language into sign language, so Deaf and hard of hearing users can receive information in the language they actually use.

“With AI Sign Language Technology, we are bridging the gap between sign language and spoken language, making real-time information and services accessible for the Deaf and hard-of-hearing communities. Our aim is to build a smarter and more inclusive world with AI,” Gong He says.

To keep the underlying Sign Language model accurate, FingerDance works with the Singapore Association for the Deaf (SADeaf) to integrate Singapore Sign Language (SgSL) into its products. The collaboration keeps Deaf and hard of hearing users involved throughout development, so their feedback shapes how the local sign language is recorded and used. The same core AI model also powers accessible sign language services for web text, news, events and conferences, extending the technology from physical settings into digital ones.

Reaching product market fit in a B2B market wasn’t straightforward. Convincing government and industry stakeholders took extensive market research and a clear demonstration of what the technology could do. Support from the NUS Enterprise ecosystem, including the NUS Graduate Research Innovation Programme (GRIP), gave the team mentorship, resources and access to industry experts that helped refine both the product and the business model.

FingerDance’s first physical deployment is SiLViA, piloted with SBS Transit at a kiosk within Chinatown MRT station. Deaf and hard of hearing commuters can say or type a query, and SiLViA responds in both spoken and sign language, including translating live audio announcements into sign language in real time.

In July 2024, FingerDance was named one of seven start ups selected for the Technology for Sustainable Social Impact (TS2) accelerator, a programme run by NUS Enterprise and the Singapore Centre for Social Enterprise (raiSE) that committed US$3.5 million to tech based social enterprises. The start-up now works out of the new BLOCK71 Social Impact Hub, a space NUS Enterprise opened for purpose driven ventures.

2023:

• Founded

2024:

• Selected (July) as one of seven start-ups for NUS Enterprise and raiSE’s Technology for Sustainable Social Impact (TS2) accelerator, which committed US$3.5 million to the cohort. Began operating out of the new BLOCK71 Social Impact Hub.

Ongoing:

• SiLViA piloted with SBS Transit at Chinatown MRT station. Working with the Singapore Association for the Deaf (SADeaf) to integrate Singapore Sign Language (SgSL) into the product with community feedback.