The impact you will make:- Design, develop, and validate signal processing algorithms for cardiac and pulmonary sounds - including filtering, noise reduction, artifact removal, segmentation, and feature extraction
>- Architect and deploy on-device machine learning models targeting resource-constrained microcontrollers, balancing accuracy against memory footprint, latency, and power consumption
>- Own the full algorithm lifecycle: from clinical requirements and dataset curation through model training, embedded deployment, and post-market performance monitoring
>- Collaborate closely with firmware engineers to integrate algorithm pipelines into real-time embedded systems
>- Partner with clinical and data science teams to define ground-truth labeling strategies, evaluate model performance on clinically relevant populations, and translate findings into algorithm improvements
>- Contribute to on-device model optimization techniques including training and deployment via frameworks such as TensorFlow Lite.
>- Develop robust offline and hardware-in-the-loop test benches using Python to validate algorithm correctness, regression-test against acoustic ground truth, and characterize edge-case behavior
>- Author and maintain algorithm documentation, design history files, and verification & validation records in compliance with FDA software development regulations and IEC 62304
>- Conduct rigorous algorithm code reviews and establish best practices for numerical stability, reproducibility, and embedded portability across the product portfolio
>
The experience you will bring:- 7+ years of experience developing and deploying signal processing algorithms in production systems
>- Strong theoretical foundation in DSP: filter design (FIR/IIR), spectral analysis and time-frequency methods (e.g. STFT)
>- Hands-on experience training and deploying ML/DL models (CNNs, RNNs, transformers) for time-series or audio classification tasks
>- Proven ability to optimize and deploy models to embedded targets - familiarity with TFLM or equivalent embedded inference frameworks
>- Proficiency in Python (NumPy, SciPy, PyTorch or TensorFlow) for algorithm development and data analysis
>- Working knowledge of C/C++ sufficient to write, review, and debug embedded algorithm implementations alongside firmware engineers
>- Experience with model compression techniques
>- Solid understanding of real-time constraints
>- Comfort working with raw physiological or acoustic signals and domain-specific data quality challenges (motion artifact, ambient noise, sensor variability)
>- Degree in a related field preferred but not required
>
It would be great to also have:- Experience in an FDA-regulated medical device environment (design controls, IEC 62304, 510(k) process)
>- Background in cardiac or respiratory audio signal processing - PCG, ECG, lung sounds, or related biosignals
>- Familiarity with clinical validation methodology: sensitivity/specificity analysis, ROC curves etc. for algorithm performance claims
>- Familiarity with MLOps practices for embedded targets: model versioning, automated retraining pipelines, and over-the-air model update strategies
>- Previous work in a startup environment where you have worn multiple hats and iterated rapidly from prototype to clinical-grade product
>- Publications, patents, or open-source contributions in signal processing, biomedical ML, or embedded AI are a plus
>- Familiarity with coding assistant tools like Claude Code for accelerated development
>
Benefits and Perks We Offer:- The opportunity to work on products that impact the health of millions of people.
- Generous paid-time off
- Stock incentive plans
- Medical/Dental/Vision, Disability + Life Insurance
- One Medical membership
- Parental Leave
- 401k Matching
- Learning and Development stipend
Work Experience & Location Requirements- This role is based out of our office in Emeryville, CA
- This is a Hybrid role (In-Office days are Tuesday, Wednesday and Thursday)
- Prolonged periods of sitting and working at a computer
- Must be able to perform job duties with or without reasonable accommodation
$171,300 - $191,400 a year
Compensation for this role is determined using a tiered, market-based compensation structure that reflects the cost of labor across different U.S. geographic locations. Salary ranges are established based on geographic labor market data and may vary depending on the employee's work location, as well as job-related knowledge, skills, qualifications, and experience.
