Head of Hardware

About The Subvocal Company

Talk to your computer without talking

Tech

At a high level, we are trying to detect the incredibly subtle physiological changes that happen when someone forms words internally, and turn those signals into continuous language. The broader sensing design space includes RF sensing, EMG, EEG, mmWave, and other non-invasive methods. For competitive and IP reasons, we are not publicly disclosing the exact architecture of our current system yet, although we can share much more during the interview process. The hard part is not just getting a model to work on one person in one recording session. These signals can change when the device moves slightly, when the same person comes back the next day, or when you move to someone with completely different anatomy. We need the system to work across people, devices, and environments, then adapt to a new user from only a few minutes of calibration. Solving that involves a mix of signal processing, self-supervised learning, large temporal models, personalization, and language decoding. We currently experiment with architectures including Conformers, Mamba-style sequence models, and pretrained speech and language models. Most of our ML stack is built in Python and PyTorch, with custom infrastructure for data collection, signal processing, distributed training, and evaluation. At the same time, we are taking a research system built from laboratory equipment and turning it into a wearable. That means building custom sensing and mixed-signal electronics, embedded and FPGA systems, custom ASICs, high-speed data acquisition, firmware, power systems, and eventually fitting everything into a small wearable. Sensor placement, electrical design, mechanical design, and the model are all tightly connected. Moving something by a few millimeters or changing part of the electronics can alter the data distribution enough to affect the model. A large part of our next phase is also a data problem. We are building the infrastructure to collect thousands of hours of high-quality subvocal speech data across thousands of people, train models continuously as that dataset grows, and understand exactly how performance scales with more users and more variation. This is what makes the work unusually interesting. There is no established playbook, and the important breakthroughs can come from a better model, a better sensing configuration, a clever piece of hardware, or simply discovering that we have been framing the problem incorrectly. The people joining now will have room to work across those boundaries and make decisions that directly determine whether the system works.

The role

We are looking for someone to take Subvocal from a working research system built with laboratory equipment to a reliable wearable that someone can use in the real world.This is a systems role. Our hardware combines sensitive physiological sensing, mixed-signal electronics, embedded and FPGA systems, high-speed data acquisition, wireless communication, batteries, firmware, mechanical constraints, and eventually custom ASICs. All of those pieces have to work together inside a device that is small, comfortable, stable, and manufacturable.The difficult part is that this is not ordinary consumer electronics. The hardware is part of the measurement itself. Changing a cable, ground plane, enclosure, sensor position, or power system can alter the data distribution seen by the model. You will work closely with the sensing and ML leads to preserve signal quality while aggressively reducing size, power consumption, and complexity.Some of the problems you will own include:* Designing custom mixed-signal, sensing, compute, power, and communications electronics.* Integrating ADCs, DACs, FPGAs, MCUs, sensor front ends, memory, and wireless systems.* Building portable prototypes that are reliable enough for large-scale data collection.* Miniaturizing the system into a small wearable and eventually a pair of glasses.* Designing for signal integrity, low noise, thermal performance, battery life, and electromagnetic compatibility.* Writing or overseeing embedded firmware, acquisition pipelines, device control, logging, and update systems.* Working with industrial designers and mechanical engineers on enclosure, fit, charging, and serviceability.* Owning component selection, BOM, prototype builds, DFM, DFT, regulatory preparation, contract manufacturers, and the path from EVT through production.* Building the hardware team and external manufacturing network around you.You might be a great fit if you have previously taken a difficult hardware product from an early prototype toward production, ideally in wearables, sensing, medical devices, robotics, AR/VR, or another tightly constrained system. You should be comfortable debugging boards at a bench, reading a schematic, working with firmware and FPGA engineers, making mechanical tradeoffs, and managing outside design or manufacturing partners without becoming dependent on them.We care much more about whether you have repeatedly made real hardware work than whether your previous title was senior enough. The right person will be equally comfortable deciding the overall architecture and personally finding the broken clock, noisy regulator, thermal issue, or intermittent connector that is preventing the prototype from working.You will have enormous ownership over both the Demo Day device and the architecture that eventually ships to customers.This is a full-time, in-person role in San Francisco.