Founding ML Engineer: Generalizable Subvocal Speech

Socket.dev; San Francisco (CA); On-site; USD 180,000 - 250,000; Full time

Job summary

Socket.dev in San Francisco is seeking a lead ML scientist to own the central problem of turning noisy physiological signals into continuous language. You will shape data collection, signal representation, model families, and evaluations, working closely with founders and sensing/ hardware leads.

We seek PhD-level research ability and production-quality coding speed, with the goal of building a product-ready, low-latency system that generalizes across people, sessions, and devices.

Qualifications

• Experience in deep learning for speech and time-series data.
• Strong research background with ability to publish or present results.
• Ability to write production-grade, scalable code.

Responsibilities

• Own the end-to-end ML problem from data to deployment.
• Define data collection, signal representation, model families, and evaluation.
• Scale training across thousands of hours and users.
• Ensure real-time latency with continuous running systems.

Skills

Deep learning for speech

Time-series analysis

Biosignals

Neuroscience

BCIs

Research ability

Production-grade code

Education

PhD in a relevant field

Tools

Python

PyTorch

CUDA

Distributed GPU training

Job description

Socket.dev in San Francisco is seeking a lead ML scientist to own the central problem of turning noisy physiological signals into continuous language. You will shape data collection, signal representation, model families, and evaluations, working closely with founders and sensing/ hardware leads.

We seek PhD-level research ability and production-quality coding speed, with the goal of building a product-ready, low-latency system that generalizes across people, sessions, and devices.