CDIIndia
Senior Hybrid Evaluation / Test Environment Lead (CXR VLM/LLM)
Hybride
This role focuses on independent evaluation and controlled experimentation, separate from core model development, to support evidence generation and safe iteration. Need someone with 9+ years of experience
Purpose of the role: Design, build, and operate a hybrid evaluation and test environment for CXR VLM/LLM models, enabling systematic testing of model functionality, edge cases, and performance across findings without interfering with the main development pipeline. Key capabilities Experience setting up model evaluation sandboxes or test harnesses for AI/ML systems, ideally in medical imaging. Ability to test and compare multiple CXR VLM/LLM variants across findings (e.g. pneumothorax, cardiomegaly, fractures) using consistent protocols. AWS experience and familiarity with cloud based dev environment. Familiarity with report-level evaluation, discrepancy analysis, and structured comparison between AI-generated outputs and clinician-validated references. Comfortable working at the intersection of engineering, clinical logic, and governance, feeding findings into monitoring, change management, and validation processes. Senior judgement to distinguish exploratory testing from evidence that is eligible for regulated use.
Informations de la mission :
Purpose of the role: Design, build, and operate a hybrid evaluation and test environment for CXR VLM/LLM models, enabling systematic testing of model functionality, edge cases, and performance across findings without interfering with the main development pipeline. Key capabilities Experience setting up model evaluation sandboxes or test harnesses for AI/ML systems, ideally in medical imaging. Ability to test and compare multiple CXR VLM/LLM variants across findings (e.g. pneumothorax, cardiomegaly, fractures) using consistent protocols. AWS experience and familiarity with cloud based dev environment. Familiarity with report-level evaluation, discrepancy analysis, and structured comparison between AI-generated outputs and clinician-validated references. Comfortable working at the intersection of engineering, clinical logic, and governance, feeding findings into monitoring, change management, and validation processes. Senior judgement to distinguish exploratory testing from evidence that is eligible for regulated use.
Informations de la mission :
- Début de mission : 12/07/2026
- Durée : 6 mois
- Date de publication : 03/07/2026
- Date limite de candidature : 30/11/2026
Compétences recherchées
CXR (Indispensable)VLM (Indispensable)LLM (Indispensable)AWS (Indispensable)cloud (Indispensable)AI/ML (Indispensable)medical imaging (Indispensable)model evaluation (Indispensable)test harness (Indispensable)monitoring (Indispensable)validation (Indispensable)discrepancy analysis (Indispensable)
Intéressé(e) ?
Rejoignez l'équipe et participez à nos projets ambitieux.