JioStar built a voice discovery assistant on the OpenAI API
Original title: JioStar’s chief architect on AI discovery at 500M-user scale
A small team launched the JioHotstar experience in about eight weeks, in more than 15 languages, ahead of a cricket tournament.
OpenAI published an interview with Vijay Seshadri, chief architect at JioStar, about a conversational video and cricket-stat discovery assistant in JioHotstar. JioStar says its platforms serve more than 500 million consumers and has supported as many as 72.5 million concurrent live-sports viewers. Using the OpenAI API, viewers can speak to find movies, shows, highlights, live-match moments, or cricket statistics. Seshadri describes a three-way tradeoff among accuracy, latency, and cost, and says more than 60% of users use voice.
Unconfirmed: Scale and usage figures come from JioStar via OpenAI’s customer interview.
- The assistant covers entertainment, sports moments, and cricket statistics.
- Work began in October and November for a February cricket tournament.
- More than 60% of users prefer voice over text, Seshadri says.
- Production evals use domain experts such as sports-magazine editors.
- Vendors
- OpenAI
- Products
- OpenAI APIJioHotstar
- Organizations
- JioStar
From the original article
7 min read at openai.comJioStar’s Vijay Seshadri explains AI discovery at 500M-user scale. Read his lessons on evals, guardrails, and taking AI into production.
openai.com
A short excerpt; the article and its images belong to the publisher. Summary, key facts and assessment above are AI-written from the article and can be wrong.