Job Description
You will own the wearable audio system end-to-end, from distributed microphone arrays and DSP to wake-word, beamforming, spatial audio, and open-ear output. The first device includes multi-mic capture, on-device processing (AEC, noise suppression, DOA, wake-word), and Class-D–driven speakers. Audio is central to the product experience.
What You’ll Do
- Own the audio system end-to-end — microphone selection and placement, audio acquisition architecture, DSP pipeline, wake-word, beamforming, AEC, noise suppression, spatial output.
- Drive part selection and benchmarking across the audio chain — microphones, audio DSPs and AFEs, Class-D amplifiers, speakers — with second-source planning.
- Specify the electrical and acoustic interfaces audio needs from the rest of the system, and partner with the EE lead on layout and the FW lead on runtime hosting.
- Drive the on-device DSP that runs in real time alongside the rest of the platform — algorithm design, latency and CPU budgets, training data, metric tuning.
- Validate the spatial mic array for direction-of-arrival accuracy, speech enhancement in noise, and wind robustness across realistic wear scenarios.
- Define the wake-word strategy — keyword choice, training data, on-device engine, false-accept and missed-trigger budgets — and own the experience users hit when they say it.
- Partners with the product design lead on acoustic detailing — microphone ports, mesh, wind shielding, housings, leakage.
- Partner with the reliability lead on component-level reliability for the audio chain — transducer durability, sweat and humidity resistance, AFE and amplifier qualification.
- Ship the product by the end of year and help build the audio team and our cloud-side voice processing capability.
Responsibilities
- 10+ years of audio engineering, with at least one shipped consumer product
- where you owned a multi-microphone voice system end-to-end.
- Deep expertise in voice DSP — beamforming, AEC, noise suppression, direction-of-arrival on multi-microphone arrays.
- Proven track record shipping wake-word or always-on voice activation at consumer scale, including data collection and metric tuning.
- Experience designing real-time DSP pipelines on constrained MCUs — not just plugin-based desktop work.
- Strong acoustic engineering instincts — microphone selection, port design, housing acoustics, wind, occlusion, body coupling.
- Direct experience driving audio component selection and benchmarking with second-source planning.
Skills
- Deploying audio DSP, wake-word, or speech models on dedicated low-power AI accelerators.
- Experience with leading on-device wake-word engines.
- Multi-channel audio DSPs and Class-D amplifier design.
- Spatial audio rendering — HRTF, panning, head-tracked binaural.
- Cloud STT pipelines and end-to-end voice UX latency budgets.
Submit resume to jobs@OSIengineering.com
Mario Ruiz
408.550.2800 x105
Type: Fulltime
Location: Palo Alto, CA (Hybrid Schedule)
Salary Range: $300k - $400k/y (DOE)
#J-18808-Ljbffr