Senior Staff Research Scientist, Speech Technologies

The Opportunity

Healthcare has a access problem — not enough clinicians, too many patients, and critical gaps that technology hasn't meaningfully closed. Hippocratic AI is changing that. Our safety-focused, healthcare-only LLM platform enables autonomous, clinical-grade conversations with patients at a scale no human workforce could achieve. With over $404M in funding, a $3.5B valuation, and partnerships with leading health systems, we're not iterating on existing tools — we're building a new category.

At the center of that platform is voice. As a Senior Staff Research Scientist in Speech Technologies, you'll lead the research and engineering that makes our AI not just intelligent, but genuinely conversational — accurate, fast, and trustworthy in the highest-stakes environments imaginable.

What You'll Accomplish

• You'll define the ASR foundation for healthcare's most advanced conversational AI. From architecture selection to production deployment, you'll shape a speech recognition system built from the ground up for clinical accuracy — one that sets the standard for what's possible when safety and performance aren't trade-offs.

• You'll solve speech recognition problems that don't have off-the-shelf answers. Medical terminology, diverse patient populations, real-world acoustic conditions — you'll design and validate models that perform where general-purpose ASR falls short, pushing the frontier of what conversational AI can reliably understand.

• You'll build the data infrastructure that makes breakthrough models possible. You'll architect the pipelines and curation processes for large-scale medical speech datasets, creating the training foundation that gives Hippocratic AI a durable, compounding advantage in clinical speech recognition.

• You'll bring research into the real world at meaningful scale. You'll close the gap between state-of-the-art methods and production systems — optimizing for latency, accuracy, and resource efficiency so that patients and clinicians experience the results of your work in every conversation.

Location Requirements

We believe the best ideas happen together. We're on the hunt for a great space in the Bellevue area — and when we find it, this role will be in office five days a week. In the meantime, you'll work fully remote with quarterly trips to our Menlo Park, CA headquarters to connect and collaborate with the team.

Responsibilities

• Design, develop, and iterate on data-driven ASR models for streaming and non-streaming conversational speech applications

• Research and implement state-of-the-art end-to-end speech recognition architectures tailored to the medical domain

• Train, evaluate, and optimize ASR models across accuracy, latency, and resource utilization dimensions

• Preprocess and curate large-scale speech datasets to support robust model training

• Collaborate closely with LLM, product, and clinical teams to integrate speech technologies into the broader Hippocratic AI platform

• Contribute to the team's research culture through experimentation, documentation, and knowledge sharing

Basic Qualifications

• PhD with 7+ years of hands-on ASR research and engineering experience, or a Master's degree with 10+ years of industry experience in speech recognition

• Deep experience designing and developing algorithms for accurate, efficient speech recognition in both streaming and non-streaming contexts

• Proven track record training and optimizing ASR models for production — balancing accuracy, latency, and compute constraints

• Experience preprocessing and curating large speech datasets for model training

• Strong Python and C++ programming skills

• Comfortable working in Linux/Unix command-line environments

• Clear communicator — you can translate complex technical work for cross-functional partners

Preferred Qualifications

• Hands-on experience building ASR systems from 0 to 1, including data pipelines, model architecture selection, and evaluation frameworks

• Practical experience with ESPnet, Kaldi, and PyTorch

• Experience with CUDA for GPU-accelerated training and inference

• Familiarity with leveraging LLMs to enhance speech recognition quality

• Experience with neural and end-to-end endpointer modeling

• Publications in tier-1 venues in speech recognition or NLP (Interspeech, ICASSP, ACL, etc.)

What You'll Love About Hippocratic AI

• The mission is real. Every improvement you make to speech recognition translates directly into better patient experiences — more accurate conversations, fewer errors, broader access to care for people who currently have none.

• The team is exceptional. You'll work alongside researchers and engineers from Google, Meta, Microsoft, NVIDIA, and Stanford, as well as clinicians and health system leaders who keep the work grounded in what actually matters to patients.

• The technical problems are genuinely hard. This isn't fine-tuning an existing API. It's building a purpose-built, safety-critical speech stack for one of the most complex and consequential domains in the world.

• The strategic moment is now. With $404M raised, a $3.5B valuation, and real deployments across leading health systems, Hippocratic AI has both the resources and the momentum to define this category — and the equity upside to match.