AI/ML Engineer (Part-Time) - Voice AI Engineer for iClinic AI, Inc
Summary
Part-time (~20 hrs/week) remote ML engineer at a Houston healthcare startup: build, evaluate, and ship speech/audio deep-learning models on patient phone-call data to help clinical teams catch problems early (heart-failure program). Core stack: Python, PyTorch, audio ML tools, and a FastAPI production backend.
Compensation: $60k – $60k • 0.0% – 1.0%
Part-time (~20 hours a week) · Remote · $60,000 per year (equivalent to $120,000 full-time) ** Path to Equity · US-based candidates only**
About iClinic AI
iClinic AI is a Houston-area healthcare startup founded by a practicing interventional cardiologist. We build production grade voice AI products for medical clinics and hospitals: AI agents that answer clinic phone lines, book appointments in the EHR, and route urgent calls. We are live with cardiology practices in Texas. We also develop disease specific voice AI products including heart failure, diabetes and dementia.
The Project
You'll join our clinical AI program in heart failure division of iClinic AI: using voice and speech data from patient calls to help care teams catch problems early. The data pipeline is already built and running; the work ahead is modeling, evaluation, and getting models safely into production. You will be expected to help other divisions as time permits.
You'll work with the engineer who built the pipeline and with our cardiologist founder on the clinical side.
What You'll Do
- Build and evaluate ML/DL models on speech and audio data for clinical prediction
- Improve data quality and preprocessing for real-world phone audio
- Design honest evaluations on small clinical datasets, with metrics a clinician can trust
- Ship models into the production backend (Python / FastAPI) with versioning and monitoring
- Write up results clearly for the founders and the research team
Must-Have
- Strong ML and deep learning fundamentals: you can explain why a model works, not just train one
- Hands-on PyTorch (or TensorFlow) and the Python data stack (NumPy, pandas, scikit-learn)
- Experience with audio or speech signals (librosa, torchaudio, openSMILE/Praat, or similar)
- Experience training and evaluating models on small, messy, real-world datasets
- Comfort shipping code to production, not only notebooks
- Based in the US and authorized to work in the US
Nice-to-Have
- Healthcare or clinical ML, and awareness of HIPAA and patient-data handling
- Time-series and longitudinal modeling
- Publications, Kaggle results, or open-source work in audio/speech ML
- Familiarity with FastAPI, Docker, Kubernetes, or Azure
Why Join
Work on a real clinical problem with a cardiologist founder, with a live data pipeline already in place. High ownership, real patient impact, and a direct path to a full-time role for strong performers. There is path to equity and additional involvement if project is successful.
How to Apply
Apply here with your resume and 3-4 lines on the most relevant ML/DL or audio project you've built (a link to code or a write-up helps). Shortlisted candidates do a short take-home and a call with the team.