Voice AI Goes Global — A Panel on Multilingual & Multiregional Deployment
Voice AI is being deployed at global scale — but making it actually work across languages, accents, and real-world conditions is still an unsolved engineering problem.
Join Maitreya Wagh (Bolna, YC-backed), Sudarshan Kamath (Smallest.ai, Series A), and Viktor Presber (KugelAudio, YC-backed) for an unfiltered panel on what it takes to build voice AI that holds up in production — across India's multilingual landscape, Europe's enterprise stack, low-bandwidth environments, and the next wave of multiregional deployments.
Speakers
Sudarshan Kamath — Co-Founder, Smallest.ai Sudarshan Kamath (CEO) and Akshat Mandloi (CTO) are the co-founders of Smallest AI, with 10+ years of experience each in Speech and AI. They lead a world-class team of researchers from MIT, Cambridge, and Max Planck Institute, united by a singular mission: building the future of Voice AI through asynchronous speech models that make AI voice truly indistinguishable from humans.
Series A raised from leading US investors including Seligman Ventures, Sierra Ventures, 3one4 Capital, Better Capital, and angels from a16z, OpenAI, and early Meta — this is a team with the research depth and investor conviction to define the future of voice.
Maitreya Wagh — Co-Founder, Bolna Forbes 30 Under 30 Asia (2026) and an alumnus of IIT Delhi, Bain & Company, and Probo. Cofounded Bolna in 2024, building voice AI infrastructure for India Inc.'s multilingual deployments at scale.
His work addresses the last-mile problem of conversational AI: making voice AI agents actually work across diverse Indian languages, low-bandwidth conditions, and the noisy, high-volume environments where most of India's business gets done. Bolna's early backers include General Catalyst, Blume Ventures, Orange Collective, and Y Combinator.
Viktor Presber — Co-Founder, KugelAudio 3x founder and co-founder of KugelAudio, building best-in-class voice AI models for Europe's enterprise infrastructure.
KugelAudio solves the sovereign AI gap for voice: on-premise speech models with 60ms latency, covering 30+ languages and dialects for real-world, high-volume customer environments. Customers include some of Germany's largest telecommunications providers, with funding currently unannounced.
Event schedule
6:00–6:30 — Arrival + drinks
6:30–7:30 — Panel + Q&A
7:30–8:00 — Networking with speakers and attendees
Who should attend
Founders building in or around voice AI — weighing whether to build proprietary models, stay model-agnostic, or go on-prem.
Operators deploying voice AI in production who want the unfiltered view on what's production-ready vs. still a research problem.
Investors tracking the voice AI stack and where the defensible technical moats actually sit.