Neethu Mariam Joy
Co-founder & CTO at SuperBryn
Building reliability infrastructure including evals, observability, and monitoring for enterprise Voice AI at SuperBryn.
Intro
I'm Neethu Mariam Joy, the Co-founder and CTO of SuperBryn. My career has been a journey from deep academic research—including a PhD from IIT Madras and a Postdoc at King's College London—to the front lines of the AI startup world. Currently, I'm focused on building the reliability infrastructure that enterprise Voice AI desperately needs, moving beyond simple demos to production-ready systems. I'm passionate about solving the 'incomplete loop' in AI, ensuring that voice agents don't just sound human, but actually execute tasks and update CRMs correctly. Whether I'm discussing the chaos of the startup ecosystem or drawing parallels between raising my four children and managing unpredictable AI systems, I'm always looking to connect with engineering teams and leaders who are serious about making Voice AI reliable and scalable.
Networking
What I can offer
- ›Deep research expertise in Speech AI
- ›Practical engineering insights for deploying non-hallucinating voice agents
- ›Guidance on Voice AI evaluation and observability
What I'm looking for
- ›Engineering teams building production Voice AI
- ›Enterprise leaders scaling voice agents
- ›High-level engineering talent for Head of Engineering roles
Best fit for
Focus
Current interests
Core competencies
Background
Career
Transitioned from deep academic research (PhD/Postdoc) in Speech AI to industry roles as an AI Scientist, eventually becoming a serial GenAI founder and CTO.
Education
Ph.D. in Speaker Normalization from IIT Madras (2011–2017); Master’s in Communication Engineering from BITS Pilani (2008–2010); Bachelor’s in Electronics and Communications Engineering from Amaljyothi College of Engineering (2003–2007).
Achievements
- ›First Women GenAI Founder from Kerala Startup Mission selected for Antler AIR4 cohort
- ›Top Scorer for both Bachelor’s and Master’s degrees
- ›Launched Super Evals Academy for Voice AI evaluation
- ›Successfully pivoted from enterprise search to SuperBryn
Opinions
- Reliability over Determinism: You can't make LLMs deterministic, but you can make the scaffolding reliable.
- Capability over Humanness: Solving problems faster matters more than sounding natural.
- Transcript-level success is a false metric; downstream data/actions are the true test.
- A tight feedback loop is more important than a perfect model.
Personality
Communication style
Professional yet highly relatable and witty, balancing technical authority with storytelling and humor.
Formality — 6/10
Vocabulary