Episode 4 with Dheemanth Reddy

What does it take to build AI that actually speaks like us, and why might voice, not keyboards, define AI’s next chapter? In Episode 4 of Slay Station, host Khushboo Mulani, Founder of Slay Media, sits down with Dheemanth Reddy, Co-founder and CEO at Maya Research, one of the new voices building voice AI out of India. From growing up in Andhra Pradesh, where he saw how technology left out people whose languages and accents mainstream systems couldn’t recognise, to building Maya 1, an open-source voice model, and Maya 2, ranked No. 1 for Hindi on Voice Arena, Dheemanth has spent his career at the intersection of AI, language, and real-world products. Backed by South Park Commons, he’s now working on making AI sound, understand, and respond like people, across the many languages and accents of India. At the heart of this episode is a conversation about where AI goes after the chatbot: why voice is becoming the interface, what it takes to build for India’s scale and diversity, and what it means for businesses, creators, and everyday users. One builder. One big bet on voice. One conversation about what AI sounds like next. Time Stamps 01:04 Who are you, what are you doing, and why are you such a big name in the AI space? 01:39 What languages does Maya have, and how does it outperform ElevenLabs? 03:18 Are your agents customised to each conversation, and how did you build that? 06:00 What are the differences between Western and Eastern languages, and have you accounted for emotion? 07:47 There’s anger sometimes, love sometimes. Can your model detect that? 09:53 How far is AI from empathising with us and being like a friend, and where is Maya in this? 11:25 Will our AI personality be fully developed by 2029? 14:07 How did you build Maya Research on the tech front? 20:16 After Maya One, how did you go about building on it? 22:45 Is there a company that already has a lot of native data? 25:54 Is personalised celebrity voices something Maya is going to do? 27:42 Where are you coming from, and what inspired you to start a voice company? 29:26 Did you break the phone down? 32:06 Beyond the keyboard, what else could we do just by talking to a screen? 35:27 Do you have data on how many people in the East get stuck because of language? 39:48 If I add more context, is that another token? And what do your investors look at? 42:03 How do you know a model’s answer is accurate, given AI hallucinations? 43:52 Was nativity your use case? 45:21 What have you integrated with so far? 52:37 Will Maya cover all 121 Indian languages and the smaller dialects in the next four years? 54:52 Will voice AI replace call centre jobs? 57:43 Not everyone has the luxury of following their heart. How do you see that? 1:07:51 Why did you name it Maya? 1:08:31 What was the story of that movie? 1:11:56 What’s next: robots, AR? What trends are you spotting in India vs Silicon Valley? 1:13:07 AI firms will hold a lot of data and power. What bad could happen? 1:15:06 Could voice have a use case in space exploration? 1:15:36 What’s the science behind voice? How do you decipher the sound of space? 1:16:09 Who designed your website, and how did you come up with it? 1:18:09 How do you build the best voice model? 1:20:17 How do you build a voice agent that actually works? 1:21:47 How do you evaluate voice AI? 1:23:23 What are the best use cases for voice? 1:23:46 We could already do consultations with AI software. How is voice different? 1:25:27 How does voice AI reach every language and become a global interface? 1:27:03 An article said Maya is beating the giants. Can Maya do that? 1:29:53 How did you end up at South Park Commons and raising money? 1:33:35 Are you part of all three SPC locations? 1:33:43 Why India? Has the ecosystem here helped you, or is Silicon Valley better? 1:38:11 Did you face rejections when cold emailing? 1:42:24 Why the telephone, banana leaf and props on the set? Disclaimer The views, opinions, and experiences shared in this podcast are those of the respective speakers and do not necessarily reflect the views of Slay Station or Slay Media. The conversations are intended for informational, educational, and entertainment purposes only. Any references to individuals, companies, products, or technologies are made in the context of personal experiences and discussions. Listeners are encouraged to form their own opinions and exercise independent judgment.