Job description

ABOUT PHONIC Phonic is an audio research lab focused on powering the most realistic, human-like voice AI conversations. We’ve rethought the entire stack in pursuit of this goal, from models to product, to create voice agents that feel like they truly understand you, respond emotionally, and perform agentic tasks with frontier intelligence. Our customers build voice-native AI products in industries like customer support, healthcare, logistics, and recruiting. We’ve grown ~20x since the start of this year, and we’ve raised over $30M from tier-1 VCs, including First Round Capital, Lux Capital, and Felicis. About the Team Phonic has an extremely talent-dense and close-knit team of top-tier AI researchers, international olympiad medalists, and former founders. We work with high trust and keep raising the bar on how we deliver world-class research and product experiences. Everyone takes ownership of what they do and isn’t afraid to dive headfirst into new problems. We’re fully in person in our SF office. ABOUT THE ROLE As a Research Scientist at Phonic, you'll drive original research that pushes the boundary of what's possible in voice AI. You'll own research directions end-to-end - from identifying the right problems to designing experiments, developing novel methods, and seeing results through to production impact. Your work will directly shape the models at the core of Phonic's product. You'll be working alongside a team of world-class researchers and engineers who move fast and hold a high bar. This isn't a role focused on incremental improvements, we're looking for someone with the taste to identify what matters, the rigor to pursue it correctly, and the drive to ship it. You'll be fully in-person in our SF office. WHAT YOU’LL DO - Identify high-leverage research problems across the voice AI stac: synthesis, modeling, alignment, perception, and take full ownership of driving them forward - Design and run rigorous experiments, then translate what works into shipped improvements to our models and product - Work directly with research engineers to move fast from prototype to production - Help set the scientific direction of the team and raise the bar on how we do research WHAT YOU’LL BRING - A track record of original research - you've found a real problem, developed a novel approach, and seen it through - Deep expertise in at least one of: speech synthesis, ASR, audio modeling, language modeling, or multimodal learning - Fluency in the math, probability, optimization, linear algebra, and the ability to reason about model behavior from first principles - You move fluidly between ideas and implementation; you don't just write papers, you build things - Clear, precise written and verbal communication NICE TO HAVE - Publications at top venues (NeurIPS, ICML, ICLR, Interspeech, ICASSP, ACL, etc.) - Experience with generative modeling, diffusion, flow matching, codec-based audio, or autoregressive models - Familiarity with RLHF or preference optimization - You are a former founder BENEFITS - 💸 Top-tier compensation: in order to get the best talent, we provide salary and equity that recognize your skillset - 🥗 Meals: free breakfast, lunch, and dinner provided in the office - 🩺 Healthcare: Comprehensive health, dental, and vision - 🤝 We have regular off-sites and team celebrations - 👵 401(k) – Let us help you plan for the future. We’ve got you covered.

Requirements

Must-haves

Identify high-leverage research problems across the voice AI stac: synthesis, modeling, alignment, perception, and take full ownership of driving them forward
Design and run rigorous experiments, then translate what works into shipped improvements to our models and product
Work directly with research engineers to move fast from prototype to production
Help set the scientific direction of the team and raise the bar on how we do research
A track record of original research - you've found a real problem, developed a novel approach, and seen it through
Deep expertise in at least one of: speech synthesis, ASR, audio modeling, language modeling, or multimodal learning
Fluency in the math, probability, optimization, linear algebra, and the ability to reason about model behavior from first principles
You move fluidly between ideas and implementation; you don't just write papers, you build things
Clear, precise written and verbal communication

Nice-to-haves

Publications at top venues (NeurIPS, ICML, ICLR, Interspeech, ICASSP, ACL, etc.)
Experience with generative modeling, diffusion, flow matching, codec-based audio, or autoregressive models
Familiarity with RLHF or preference optimization
You are a former founder

Application prep

Use the role’s language and application details while tailoring your materials.

Keywords to mirror

Platform: ashby
Effort: quick

Benefits

Benefits detected in the listing.

Dental

Vision

401(K)