AI Engineer, Health Informatics

You will build the models that read what is messy in medicine (clinical notes, referral letters, patient messages, voice transcripts) and turn it into structured, sourced facts on one person's timeline.

Your work feeds the rules; it never replaces them. Every output has to be checkable against the source it came from.

• Building extraction and coding models from clinical text to ICD-10, SNOMED CT, LOINC and RxNorm
• Measuring accuracy against source documents and clinician review, and publishing how
• Designing evaluation sets that reflect real hospital data, not benchmarks
• Working with the clinical knowledge engineer so model outputs meet each rule's inputs
• Writing and maintaining tests across the system

In your first 12 months: extraction in production on real clinical text with measured, published accuracy, and ownership of the learned layer.

• PhD in health informatics, biomedical informatics, computer science or a related field
• Hands-on experience with clinical text or EHR data (MIMIC, institutional extracts, OMOP or FHIR)
• Strong Python and modern NLP or LLM tooling
• Rigour about evaluation: calibration, error analysis, and knowing when a model should abstain

加分條件

• Published work in clinical NLP or health AI
• Working knowledge of clinical terminologies
• Experience deploying models under privacy constraints

We care about what you can do. A candidate with hands-on clinical data experience matters more than credentials. Classical ML and biostatistics background is a better match than deep learning for what we are building.

Chroneme is building neuro-symbolic AI for medicine: one record of each person, in order and in context, read by learned models and decided by explicit clinical rules that state their reason and their source. It tells the care team who needs attention and why, and prepares the work for a clinician to sign.

We are a small team in Birmingham, Michigan, building with clinicians and validating under a pre-registered protocol. You will work directly with the founders, with no layers between you and the work. Early team members get meaningful equity.

1. 1An introductory call about your background and interest (30 minutes, with a co-founder)
2. 2A working conversation where you walk us through a project in depth and we discuss how you would approach ours (60 minutes, with the founding team)