CalTek is an Engineering and IT staffing agency that is recruiting an Artificial Intelligence (AI) Engineer for a client. Our client is a pioneering AI healthcare technology firm specializing in FDA-cleared Software as a Medical Device (SaMD) for advanced medical imaging analysis. Their platform integrates seamlessly into hospital and radiology workflows, using AI-powered algorithms to detect and prioritize urgent, life-threatening conditions—such as ischemic strokes, pulmonary embolisms, and intracranial hemorrhages—within seconds.
With a strong track record of regulatory compliance and clinical performance, this company leads the way in AI diagnostic tools that meet the highest standards of safety, efficacy, and reliability.
Typical Duties and Responsibilities
As an AI Engineer, you will be responsible for designing and deploying AI models within the stringent technical and regulatory boundaries of FDA-cleared SaMD. You’ll collaborate with cross-functional teams to develop clinically validated algorithms, build compliant data pipelines, and support lifecycle documentation for regulatory audits and product releases.
Key Responsibilities:
- Develop and validate AI/ML models (e.g., CNNs, U-Nets, transformer-based models) for detection, classification, and segmentation of radiological abnormalities using DICOM imaging modalities (CT, MRI, X-ray)
- Build scalable training, testing, and inference pipelines using TensorFlow, PyTorch, and ONNX, ensuring reproducibility, version control, and traceability
- Perform robust clinical validation in accordance with FDA and ISO standards, including sensitivity/specificity metrics, ROC/AUC analysis, and reference standard comparison
- Collaborate with QA/RA teams to support documentation for 510(k) submissions, SaMD change control, and post-market surveillance
- Maintain compliance with IEC 62304, ISO 13485, ISO 14971, and FDA’s Good Machine Learning Practice (GMLP) guidelines
- Integrate explainable AI outputs (e.g., Grad-CAM, SHAP) to meet interpretability requirements for clinician-facing tools
- Collaborate with product and engineering teams to deploy AI models within a QMS, containerized via Docker, and integrated into PACS viewers or cloud-based DICOM routing systems
- Conduct algorithm performance monitoring and model update risk analysis as part of SaMD lifecycle management
- Support traceability between data, training code, performance metrics, and intended use cases
Education
- Bachelor’s degree in Computer Science, Biomedical Engineering, Electrical Engineering, or related discipline required
- Master’s or Ph.D. preferred, particularly with a focus on AI in healthcare, computer vision, or biomedical imaging
Required Skills and Experience
- 5+ years of hands-on experience developing AI/ML models in a regulated healthcare or med-tech environment
- Expertise in deep learning frameworks such as TensorFlow, PyTorch, and Keras
- Strong background in medical image processing, DICOM standards, and preprocessing for radiology workflows
- Familiarity with FDA SaMD development frameworks, especially related to 510(k), De Novo, and PMA submissions
- Experience working within or integrating with QMS platforms and maintaining compliance with ISO 13485 & IEC 62304
- Skilled in Python programming, data analysis, model debugging, and algorithm optimization
- Knowledge of regulatory guidelines for software lifecycle validation, including design input/output traceability, test protocols, and risk analysis
- Hands-on experience with MLOps tools, model versioning systems (e.g., MLflow, DVC), and secure cloud platforms (AWS, Azure, or GCP)
- Excellent documentation and communication skills—capable of writing clear, audit-ready technical files and regulatory responses
Preferred Qualifications
- Prior work on FDA-cleared AI/ML medical devices or submissions under the 510(k) or De Novo pathways
- Experience with algorithm change management under locked or adaptive ML models
- Familiarity with cybersecurity best practices for cloud-deployed SaMD (FDA premarket cybersecurity guidance, NIST CSF)
- Experience in multi-center clinical validation studies, reference reader workflows, or ground-truth adjudication
- Understanding of AI risk management per ISO/TR 24028 and GMLP principles
- Contributions to peer-reviewed publications or open-source medical AI projects
