Machine Learning Engineer; Model Development

Kadence Talent is partnered with a pre-Series A healthtech AI company in the Bay Area, looking for a ML Engineer (Data Scientist, model development)

Medical coding is a high-stakes domain: a model error isn't just an incorrect prediction — it can become a billing, compliance, revenue, or trust issue. We're looking for someone who can improve model quality systematically, combining strong ML engineering, evaluation design, prompt engineering, and domain learning.

We're hiring an ML Engineer to build and improve specialty-specific medical coding agents. You'll work closely with engineering, clinical coding experts, auditors, and customers to turn model failures into measurable improvements.

This is a hands-on role at the intersection of applied machine learning, LLM-based agent development, evaluation systems, and healthcare data workflows. You'll own the feedback loop that takes audited coding decisions, clinical edge cases, and production failures — and turns them into better agents.

Build, maintain, and iterate on specialty-specific coding agents. Translate expert and auditor feedback into concrete improvements. Run structured improvement cycles — batched feedback, test sets, regression checks — rather than one-off fixes.

Build evaluation harnesses and gold-standard datasets. Create regression tests so changes don't silently break existing behavior. Build confidence scoring and metrics that are meaningful to both engineers and clinicians.

Design prompts for clinical evidence extraction, coding recommendations, and rationale generation. Build and improve agent scaffolding — prompt chains, tool use, evaluation loops. Work toward model-agnostic evaluation across foundation models.

Partner directly with certified medical coders to understand coding rules and edge cases. Convert their feedback into precise engineering tasks. Help non-technical stakeholders understand model behavior and tradeoffs.

• 2+ years building applied ML, LLM-based systems, or AI agents in production
• Strong Python skills and comfort with messy, real-world data
• Experience building evaluation frameworks, benchmarks, or quality-measurement systems
• Solid statistical reasoning — precision/recall, confidence intervals, error analysis
• Hands-on experience with LLMs, prompt engineering, or agentic systems
• Ability to debug model failures by reading examples closely and designing structured fixes
• Comfort with early-stage ambiguity and fast iteration

Healthcare, claims, or clinical NLP experience; familiarity with CPT/ICD-10/payer rules; human-in-the-loop ML systems; document AI/OCR;
Claude/OpenAI API experience; HIPAA/SOC 2 exposure.