Research Scientist, Speech Technologies (Senior, Staff, Senior Staff)

ROLE MISSION As Research Scientist in Speech Technologies, you will lead the research and engineering that makes Hippocratic AI's conversational platform not just intelligent, but genuinely conversational—accurate, fast, and trustworthy in the highest-stakes environments imaginable. You'll define the ASR foundation for healthcare's most advanced conversational AI, ensuring every patient interaction is understood with clinical precision. This role exists because accurate speech recognition in clinical contexts is a frontier problem—no off-the-shelf solution exists, and the work directly determines whether our platform can reliably serve the millions of patients who need it. WHAT YOU WILL ACCOMPLISH Own your first major outcome: By day 90, you will have shipped measurable improvements to our production ASR system (improved accuracy on medical terminology, reduced latency, or expanded robustness to diverse patient populations), validated performance gains on clinically relevant benchmarks, and established the data infrastructure roadmap that will compound our advantage in medical speech recognition. Drive lasting impact: At 12 months, you will have designed and deployed a next-generation ASR architecture purpose-built for healthcare, built the large-scale medical speech dataset pipeline that gives Hippocratic AI durable competitive advantage, published your research at tier-1 venues, and directly shaped how millions of patients experience conversational AI through your speech technology innovations. THE TEAM You'll lead a team of researchers and engineers building speech technologies that matter. You'll work alongside ML researchers, software engineers, and clinicians from Google, Meta, Microsoft, NVIDIA, and Stanford—as well as health system leaders who keep the work grounded in real clinical needs. This is a culture of rigorous research, rapid iteration, and solving genuinely hard technical problems that have real-world impact. WHAT YOU'LL DO - Design and develop data-driven ASR models for both streaming and non-streaming conversational speech applications, architecting end-to-end speech recognition systems purpose-built for medical accuracy, latency, and robustness - Research and implement state-of-the-art speech recognition architectures tailored to the medical domain, addressing problems that off-the-shelf ASR cannot solve—medical terminology, diverse patient populations, real-world acoustic conditions - Train, evaluate, and optimize ASR models across accuracy, latency, and resource utilization—balancing clinical precision with production constraints to ensure seamless integration into our platform - Build data infrastructure and curation pipelines for large-scale medical speech datasets, architecting the training foundations that create durable, compounding advantages in clinical speech recognition - Collaborate with LLM, product, and clinical teams to integrate speech technologies into the broader Hippocratic AI platform, translating patient and clinician feedback into research priorities - Contribute to research culture through rigorous experimentation, documentation, publications at tier-1 venues, and knowledge sharing that elevates the team's technical capabilities Location Requirement We believe the best ideas happen together. This role is based in our Menlo Park, California office, expected to be five days a week. We're also exploring establishing a presence in the Bellevue area—if that develops, flexibility on location may be available for exceptional candidates. WHAT YOU BRING MUST‑HAVE: - PhD with 3+ years of experience in Speech Recognition or related field or Masters with 5+ years of hands on experience with ASR. - Experience Designing and developing algorithms for accurate and efficient speech recognition for both Streaming and Non-Streaming use cases. - Experience with Training, evaluating, and optimizing ASR models for various factors including accuracy, latency, and resource utilization. - Experience with Preprocessing and curating large speech datasets for training models. - Strong programming skills with working knowledge of Python & C++ - Comfort working in a Linux/ Unix command-line environment. - Team player with good communication skills (oral and written) NICE‑TO‑HAVE - Experience with building 0 to 1 ASR solutions, including setting up data pipelines, SOTA model architectures and evaluation pipelines. - Hands-on Experience with ESPNET, Kaldi and Pytorch. - Experience with CUDA. - Experience with leveraging LLMs for enhanced speech recognition tasks. - Experience with Neural/ E2E Endpointer modeling. - Publications in tier 1 journals in the field of speech recognition/ NLP. COMPENSATION Compensation is based on experience, expertise, and level of responsibility. We offer competitive packages that reflect the seniority and scope of the role, along with equity, health insurance, and other benefits. WHY JO