The RoleAs a Senior Machine Learning Engineer, you will build the intelligence layer that automates complex healthcare compliance and document procurement workflows. You will own systems that turn noisy, unstructured inputs such as faxes, phone transcripts, and operational data into reliable structured facts, decisions, and downstream actions.
This is not a pure research role. It is a product and systems role for someone who knows how to turn modern foundation models into dependable production infrastructure. You should be excited by messy real-world data, ambiguous edge cases, and high-leverage workflow automation. You will work across LLMs, OCR pipelines, voice AI, evaluation systems, and backend production infrastructure to help automate the DME process end to end.
What You'll Work On
- Build and improve multimodal AI pipelines that process healthcare documents, OCR output, transcripts, and workflow context into structured facts and decisions.
- Design LLM-powered extraction, classification, validation, and routing systems for operational and clinical workflows.
- Improve document intelligence systems across OCR, schema extraction, confidence scoring, error handling, and low-quality input recovery.
- Develop voice AI workflows for patient and provider outreach, transcript understanding, post-call extraction, and follow-up automation.
- Create evaluation harnesses, benchmarks, and regression tests for extraction quality, hallucination prevention, workflow accuracy, and model changes.
- Decide when to use LLMs, deterministic logic, retrieval, human review, or hybrid systems to maximize quality and reliability.
- Partner with product and engineering to identify the highest-leverage automation opportunities and translate them into shipped systems.
- Optimize cost, latency, and reliability across model providers and infrastructure layers.
- Work closely with backend engineers to deploy AI systems into our AWS and serverless environment with strong observability and operational rigor.
Technical Requirements- Strong experience building production AI systems around LLMs, OCR, and unstructured data workflows.
- Proven track record shipping applied AI products, not just prototyping models offline.
- Deep familiarity with modern LLM workflows including prompting, structured outputs, tool use, retries, fallbacks, guardrails, and model evaluation.
- Experience with document intelligence systems such as OCR pipelines, document extraction, classification, post-processing, and confidence-based review flows.
- Experience with voice or conversational AI, or adjacent systems involving transcripts, call automation, and conversational extraction.
- Strong proficiency in Python and comfort working in production codebases with APIs, queues, and backend services.
- Experience deploying and operating AI systems in AWS or similar cloud environments, including serverless or event-driven architectures.
- Strong instincts around evaluation, benchmarking, monitoring, and quality assurance for real-world AI systems.
- Ability to work across structured and unstructured data and design systems that are robust to noisy, incomplete, and ambiguous inputs.
Nice to Have- Experience in healthcare, claims, revenue cycle, or regulated operational environments.
- Experience with human-in-the-loop workflow design and review tooling.
- Familiarity with telephony vendors, speech systems, or conversational agent infrastructure.
- Experience comparing and routing across model providers such as OpenAI, Anthropic, Bedrock, or equivalent.
- Experience designing internal tools or operational systems used directly by workflow teams.
- Background in machine learning, applied NLP, information extraction, or related fields.
