Krish Ramineni, Fireflies CEO and Microsoft alum, learned the value of NLP working with Skype and Office as a Product Manager. He set out to solve a problem he had: note-taking in meetings and following up afterward. Fireflies has been used by more than 60,000 organizations to make meetings more efficient. Krish has raised nearly $20M from an A-list group of investors including Canaan Partners and Khosla Ventures.
Listen and learn...
1. The evolution of speech recognition technology in the enterprise 2. How Krish and the team build an AI voice assistant that joins meetings in 100 countries every day 3. How to start with 85% ASR (automated speech recognition) accuracy and make it better using AI 4. How to mitigate the impact of biased training data where foreign accents and uncommon speech patterns are underrepresented 5. Who owns voice transcripts used to train AI models 6. How being recorded changes participant behavior in meetings 7. The future of "voice-first" computing
References in this episode:
Krishna Gade from Fiddler discusses AI explainabilityFireflies • on Twitter A whole breast, AI-based ultrasound system was cleared by the FDA to improve mammogram accuracyDentists now use AI to improve the accuracy of spotting cavities
Krish Ramineni, Fireflies CEO and Microsoft alum, learned the value of NLP working with Skype and Office as a Product Manager. He set out to solve a problem he had: note-taking in meetings and following up afterward. Fireflies has been used by more than 60,000 organizations to make meetings more efficient. Krish has raised nearly $20M from an A-list group of investors including Canaan Partners and Khosla Ventures.
Listen and learn...
1. The evolution of speech recognition technology in the enterprise 2. How Krish and the team build an AI voice assistant that joins meetings in 100 countries every day 3. How to start with 85% ASR (automated speech recognition) accuracy and make it better using AI 4. How to mitigate the impact of biased training data where foreign accents and uncommon speech patterns are underrepresented 5. Who owns voice transcripts used to train AI models 6. How being recorded changes participant behavior in meetings 7. The future of "voice-first" computing
References in this episode:
Krishna Gade from Fiddler discusses AI explainabilityFireflies • on Twitter A whole breast, AI-based ultrasound system was cleared by the FDA to improve mammogram accuracyDentists now use AI to improve the accuracy of spotting cavities
Nyd den ubegrænsede adgang til tusindvis af spændende e- og lydbøger - helt gratis
Dansk
Danmark