ML Engineer — Speech + Vision (Contract)
COMPANY code:blue
LOCATION Remote
JOB TYPE Contract / Part-time
DESCRIPTION
We're building an AI system that detects early signs of stroke using everyday devices. No special hardware. No clinic. Just a phone, a face, and a few seconds of speech — analyzed in real time.
This isn't a demo. It's a clinical-grade tool we're preparing for our next human study.
What you'll work on:
- Improve speech slurring detection (Wav2Vec2 / Whisper / SpeechBrain)
- Benchmark and harden our MediaPipe-based facial droop pipeline
- Build and deploy cloud inference on AWS (SageMaker, Lambda, S3)
- Define clinical-grade evaluation metrics — sensitivity, specificity, real-world noise tolerance
This is 6-8 weeks of real work. We start with a paid 2-week trial sprint so both sides can assess fit — then extend through our study prep if it's working.
What we're looking for:
You've shipped production ML at Meta, Amazon, Google, Apple, Microsoft — or a top AI research lab. You know what good looks like. You're not looking for full-time right now — but you want your side work to mean something.
Stack: Python · PyTorch / TensorFlow · AWS · MediaPipe · Wav2Vec2 / Whisper
Compensation: Competitive contract rate — commensurate with senior ML experience.
Apply here or DM me directly — I read every message.