Medallion is a leading provider operations platform focused on improving healthcare team efficiency. The Staff ML Engineer will design and optimize AI systems to automate credentialing and provider verification processes, working autonomously to deliver impactful solutions.
Responsibilities:
- Scope and lead ML initiatives end-to-end from identifying opportunities and defining the problem through production deployment and iteration
- Design, develop, and optimize ML models and AI systems for document parsing, extraction, classification, and intelligent automation
- Build and maintain production ML pipelines that are robust, observable, and scalable
- Integrate and fine-tune third-party AI services (OpenAI, Amazon Textract, cloud ML APIs), managing cost, latency, and quality tradeoffs
- Analyze datasets to uncover patterns, validate model performance, and generate actionable insights
- Help develop our AI roadmap, balancing key technical and product tradeoffs
- Drive architectural decisions for ML systems and establish best practices for development, evaluation, and deployment
- Teach and mentor members of the engineering team, constantly modeling how great ML software should be developed
Requirements:
- 8+ years of experience as a software engineer, with 4+ years focused on ML or applied AI in production environments
- Track record of shipping ML systems that deliver measurable business impact
- Strong proficiency in Python and the modern ML/AI stack (PyTorch, Hugging Face, scikit-learn, or similar)
- Experience with LLMs in production including fine-tuning, prompt engineering, RAG, and evaluation strategies
- Strong ability to work cross-functionally to help define, build, and deliver on product and tech objectives
- Experience mentoring and leading teams, ideally in a startup environment
- Care deeply about both technical success and product success
- Excellent communication skills
- Experience with healthcare data, compliance workflows, or regulated industries
- Background in document understanding, OCR/ICR, or information extraction from unstructured data
- Hands-on experience with MLOps tooling (MLflow, W&B, Airflow) or cloud ML infrastructure (SageMaker, Vertex AI, Bedrock)
- Graduate work in a quantitative field (CS, statistics, ML, or related)