Lead Data & AI Engineer (Clinical EMR)

Team: Product & Engineering (Data/AI)

Reports to: Head of Product (dotted line to Head of Engineering)

Type: Full-time

Role Purpose

Build, deploy, and maintain the AI/ML stack powering our EMRclinical NLP/LLM, decision support, and voice scribing. Own end-to-end data engineering, model training, and MLOps with healthcare compliance baked in (PDPA, HIPAA, ISO 27799).

What Youll Do

1. Architect and operate pipelines for structured/unstructured clinical data (EHR notes, HL7 v2, FHIR, audio).

1. Build/maintain the feature store for clinical AI (labs, meds, allergies, vitals, orders) with lineage & versioning.

1. Implement PHI de-identification/re-identification, KMS-backed encryption, DUAs, and access controls.

1. Own the extraction + normalization stack for: problems/diagnoses, symptoms/findings, medications (with attributes), labs, orders, allergies.

1. Ship a hybrid extractor (transformer NER + rules) with assertion

(present/absent/etc.) and temporality.

1. Build a medication attribute parser (dose, unit/UCUM, route, frequency, duration, PRN, instructions).

1. Implement a two-stage entity linker (candidate gen via lexicon/vector search + cross-encoder rerank) to SNOMED CT, RxNorm, LOINC; manage crosswalks to ICD-10/CPT and local catalogs.

1. Operate ontology ops: version pinning, diffs, UMLS/SNOMED licensing, regression tests per ontology release.

1. Enforce safety guards (drugallergy, duplicate therapy, dose range) and confidence-driven UI disambiguation.

1. Build RAG/LLM pipelines for summarization, CDS, and scribe workflows (prompting, tool use, retrieval, guardrails).

1. Integrate ASR + diarization with streaming partials/hotwords for clinical terms and names.

1. Stand up MLOps: model registry, experiment tracking, CI/CD, canary/shadow deploys, drift & safety monitoring, blue/green rollbacks.

1. Meet SLOs: p95 speechdraft less than 2.0s, ASR partial updates every 300500ms, 99.9% uptime, rollback

   less than

   5 min.

1. Optimize inference (Triton/ONNX Runtime, quantization/distillation, caching) and track cost per encounter.

1. Ship SMART on FHIR apps and CDS Hooks; design gRPC/REST services; run Kafka/PubSub with idempotent consumers.

1. PHI-safe prompts/logs, prompt-injection & data-exfiltration guards, constrained tool allowlists.

1. Audit trails exportable for clinical review & compliance.

1. Partner with Product & Clinical to encode guidelines/rules alongside ML.

1. Mentor engineers; uphold code quality, reviews, and on-call.

Required Qualifications

1. 4+ years Data/ML Engineering (healthcare strongly preferred).

1. Expert: Python, SQL, PyTorch/TensorFlow, Hugging Face.

1. Deep NLP/LLM (transformers, RAG, prompt engineering, guardrails).

1. Standards: FHIR/HL7, SNOMED CT, ICD-10, RxNorm, LOINC, CPT; UMLS

familiarity.

1. MLOps (MLflow/Kubeflow/Vertex/SageMaker), containerized inference, CI/CD.

1. Privacy/security in regulated environments (PDPA/HIPAA/ISO 27799).

Nice to Have

1. ASR/diarization (Whisper, Vosk, Kaldi), ONNX/TensorRT, Triton; gRPC/WebRTC streaming.

1. GPU scheduling, vector DBs, OpenTelemetry, Terraform/IaC.

Success Metrics (you own)

1. NER micro-F1 0.92 (per-type 0.88).

1. Linking top-1: SNOMED/RxNorm/LOINC 0.95 (top-5 0.99).

1. Med attributes exact-match 0.93; UCUM unit validity 0.99.

1. Safety: drugallergy recall greater than 99%, precision greater than 95%.

1. Latency/Reliability: p95 speechdraft less than 2.0s; streaming extraction p95

   less than

200ms/chunk; 99.9% uptime.

1. ASR clinical WER 12%; partial stability 0.90.

1. Cost per encounter within target.