We are looking for Senior MLOps specialist. It will be full-time for about 3 months. We are flexible on location and rate for the right person.

Mission

• Build and run a scalable, traceable, FDA-ready data platform bridging on-prem DGX and AWS for our 2026 submission.

Key responsibilities

• Implement ingestion + QC automation starting with CT DICOM, expanding to video/C-Arm and radiology reports.

• Implement/operate distributed processing on AWS Batch (Spot) for large-scale QC + predictions + derivatives.

• Deploy ClearML for dataset versioning/lineage and experiment tracking (runs/metrics/artifacts; provenance).

• Optimize training data access using Lance (or equivalent) for fast loading and incremental updates.

• Build a PostgreSQL-backed enrichment service for metadata/labels/predictions independent of raw media (optional search via OpenSearch/text+vector).

• Integrate/operate labeling workflows (Encord preferred; alternatives acceptable), incl. RBAC/QC/audit trail + algorithmic label ingestion.

• Establish a governed clinical validation environment meeting 21 CFR Part 11 expectations (access control, audit trail/WORM, provenance) and HIPAA/PHI handling.

Requirements

Python, AWS (S3, AWS Batch, EC2 Spot), Kubernetes, HPA, Karpenter, Networking (VPC, VPN UiFI), Trivy, Jenkins, PostgreSQL, DICOM, PACS, Orthanc, ClearML, Lance, Encord, OpenSearch, FFmpeg

Knowledge: Medical imaging (CT, DICOM), data ingestion pipelines, data quality (QC) automation, distributed data processing, experiment tracking and dataset versioning, metadata and label management, clinical AI validation, 21 CFR Part 11 compliance, HIPAA/PHI handling, RBAC, audit trails, WORM storage, vector search, media processing, annotation/labeling workflows, provenance and data lineage, PACS integration, radiology workflows.

Benefits