speech-ai
(12)
Anant Singh and the Business of Being Understood
He sells software that changes how a voice is heard, while insisting the person behind it must remain intact. At Sanas, Anant Singh has made that tension the center of a global sales story.
Madhav Lavakare Spent Eight Years Putting Conversation Back in View
A school friend’s unanswered problem sent a 16-year-old to the family projector with a screwdriver. Eight years, two gap years and a patient procession of prototypes later, AirCaps finally shipped.
Prashant Shishodia Is Teaching Machines to Hear the Whole Conversation
From a scholarship classroom in India to Google-scale machine learning and a two-person San Francisco audio lab, Shishodia has kept returning to the same useful obsession: make ambitious models efficient enough to matter.
Retell AI
Meet your AI call center from the future.
Dialogus
Dialogus is a San Francisco startup (Y Combinator S26) building infrastructure for enterprise voice agents. Instead of automating a single call, it aims to rebuild the contact center itself: AI voice agents that answer live phone calls, resolve customer requests end-to-end, execute workflows against CRMs and internal systems, hand off to humans when needed, and produce a full audit trail. The platform is SOC 2 compliant and, per the company, already handles thousands of production calls for large enterprises including Papa Johns, KFC and Totalplay.
Vu Van
Vu Van is the Vietnamese-born co-founder and CEO of ELSA (English Language Speech Assistant), an AI-powered app that gives non-native speakers instant feedback on English pronunciation, fluency and intonation. She built the idea out of her own struggle to be understood as a Stanford graduate student, teamed up with speech-recognition scientist Xavier Anguera, and has grown ELSA into one of the world's largest AI language platforms, with tens of millions of users across roughly 190 countries. She is a World Economic Forum Technology Pioneer and Endeavor entrepreneur.
Dylan Fox
Dylan Fox is the founder and CEO of AssemblyAI, an AI research company building Speech AI models that transcribe and understand human speech through a developer-facing API. A former machine learning engineer at Cisco, he started AssemblyAI as a solo founder in 2017, went through Y Combinator, and grew it into a platform used by companies like Spotify, Zoom, and Fireflies.ai. The company processes tens of millions of API calls a day and has raised roughly $115M, including a $50M Series C led by Accel in December 2023.
Flair Labs
Flair Labs builds lifelike voice AI 'digital workers' for mortgage and real estate teams. Spun out of Stanford's AI lab and backed by Y Combinator, its agents answer, qualify, follow up with, and book borrowers across voice, SMS, and email around the clock, aiming to handle the repetitive conversational work that keeps loan officers off the phones and let humans focus on judgment, empathy, and expertise.
Marty Sarim
Marty Massih Sarim is the President of Sanas, a real-time speech AI platform that modulates accents and eliminates background noise for contact center agents globally. An Afghan refugee who came to the US in the early 1980s and started his career at 18 as a call center agent, Sarim brings over 25 years of BPO and contact center industry expertise to Sanas. He is also Co-Founder and General Partner at Carya Venture Partners, a $20M micro-fund focused on deep tech and enterprise AI, and founder of the Moe123 Scholarship Fund, which has awarded $125,000+ to high school seniors in the Minneapolis area.
Maxim Serebryakov
Maxim Serebryakov is the Co-Founder and CEO of Sanas, a Palo Alto-based AI company building the world's first real-time Speech Understanding Platform. Born in New York and raised in Russia, Serebryakov co-founded Sanas at Stanford with two fellow international students after witnessing a friend face accent discrimination in a contact center. The company's AI modulates accents in real time while preserving a speaker's voice, tone, and emotion - serving 150,000+ live agents across 39 countries. Sanas has raised $149.2M in total funding, including a $65M Series B led by Quadrille Capital in February 2025.
Sanas
Sanas builds real-time speech understanding AI for contact centers - accent translation, language translation, noise cancellation and speech enhancement that runs live on a call without losing the speaker's voice or emotion.
Ashish Nagar
Ashish Nagar is the Founder and CEO of Level AI, a Mountain View-based enterprise AI company that has raised $73.1 million to transform how contact centers operate. Armed with a B.Tech in Applied Physics from IIT Delhi, a Stanford MS, and a Stanford GSB MBA, Nagar cut his AI teeth building the Alexa Prize at Amazon — a project to make Alexa hold a 20-minute conversation on any topic, collaborating with researchers from MIT, CMU, Stanford, and Oxford. He founded Level AI in 2019 after recognizing that the same ambient AI breakthroughs powering voice assistants could be turned toward the unglamorous but massive world of customer service. The company's vertically integrated, CX-native large language model now analyzes 100% of customer conversations for enterprises like Affirm, Penske, and Carta, detecting seven distinct emotions, cutting call handling time by 10-25%, and onboarding agents 30-50% faster.