AI Research Scientist (Audio/Voice) (Hayward)

Artificial Intelligence Research ScientistRedwood City, CA (on-site M-F)Our client is a high-growth AI company building next-generation voice agents for customer interactions across sales, support, and operations. Backed by leading investors, the team is scaling rapidly with strong revenue traction and a focus on real-time, production-grade AI systems.About the RoleThis is a research-driven, high-impact role for ML researchers who want to push the boundaries of real-time conversational AI. You'll focus on advancing model capabilities for human-like voice agents operating in complex, real-world environments, spanning LLMs, speech, and multimodal systems. Your work will bridge cutting-edge research with deployment, directly shaping core product experiences and infrastructure.ResponsibilitiesResearch and develop new techniques across LLMs and audio models to improve reasoning, latency, and conversational quality in real-time voice systems.Rapidly prototype, train, and iterate on experimental models and pipelines, turning research ideas into working end-to-end prototypes.Design and maintain evaluation frameworks, datasets, and metrics to benchmark performance on complex, real-world voice and dialog tasks.Collaborate with engineering to translate research insights into scalable, production-ready systems powering live customer conversations.Build human feedback loops and annotation workflows to incorporate subjective conversational quality signals into model improvement.Stay at the frontier of ML research and bring new approaches into the stack, from pre-training and post-training methods to multimodal architectures for voice agents.Qualifications1–5 years of experience in audio and/or multimodal ML research or engineering in industry or academia.Strong ML research background in areas such as LLM pre-training/post-training, ASR, TTS, or multimodal systems.Deep technical foundation in modern ML, including PyTorch, model architectures, and the underlying math.Proven ability to design and run experiments on open-ended problems, iterate quickly, and analyze complex model behavior.Track record of translating research into working systems in real-world, latency-sensitive environments.Excellent communication skills and comfort working cross-functionally in fast-paced, high-ownership settings.Preferred SkillsAdvanced degree (MS or PhD) in Computer Science, Machine Learning, AI, or a related field, or equivalent research-level experience.Experience shipping models for real-time voice or conversational products (for example, call centers, assistants, interactive agents).Prior work with reinforcement learning from human feedback (RLHF), preference modeling, or large-scale human evaluation pipelines.Experience in early-stage startups or small, high-performing ML teams with high ownership ceilings.