Lead ML Architect: Generalizable Signal Decoding

Empathy Talent; San Francisco (CA); Hybrid; USD 120,000 - 200,000; Full time

Job summary

Empathy Talent in San Francisco, CA is seeking a Machine Learning Engineer to own a core technical challenge: turning weak, noisy physiological signals into continuous language, with generalization across users and devices.

You will define data collection and representation, explore model families, design evaluation protocols, and scale training across thousands of hours and users, delivering real‑world, low‑latency inference.

Qualifications

• Deep experience applying ML to noisy signal data with distribution shifts.
• PhD-level research ability or equivalent demonstrated by publications.
• Ability to write production-quality code and run independent experiments.

Responsibilities

• Develop generalizable representations from large unlabeled time-series data.
• Build sequence models using Conformers, Mamba-style architectures, CTC, transducers.
• Adapt cross-user models with ~15 minutes of calibration data.
• Design robust evaluation frameworks across held-out sets and devices.
• Scale training across thousands of hours and users.

Skills

Time-series modeling

Production-quality code

Experiment design

Independent researcher

Cross-disciplinary collaboration

Education

PhD-level research ability

Tools

Python

PyTorch

TensorFlow

Experiment tracking

Job description

Empathy Talent in San Francisco, CA is seeking a Machine Learning Engineer to own a core technical challenge: turning weak, noisy physiological signals into continuous language, with generalization across users and devices.

You will define data collection and representation, explore model families, design evaluation protocols, and scale training across thousands of hours and users, delivering real‑world, low‑latency inference.