Knowledge integration
Connected venue-specific knowledge retrieval with the language model and kiosk interface.
VirtualConcierge
A voice and touchscreen concierge grounded in site-specific knowledge, with an asynchronous interaction pipeline.
My contribution
Voice workflow
Worked on speech output and an asynchronous interaction pipeline.
Integration reliability
Diagnosed retrieval and voice issues, then supported field validation.
System and engineering notes
System flow
Voice or touch input moves through venue-specific retrieval, a language model response, and speech or interface output.
Engineering focus
The integration work focused on keeping the interaction pipeline asynchronous and the kiosk interface responsive while diagnosing retrieval and voice failures.
Reliability evidence
Field work included repeated-recognition and silent-audio fixes, followed by recorded in-person kiosk acceptance in August 2026.
What was field-tested
- Site-specific knowledge retrieval
- Voice and touchscreen interaction
- Recorded in-person kiosk acceptance, August 2026
Have a similar problem to solve?
Get in touch