AI Solutions Engineer

AI Solution Engineer, Clinical Imaging
Lumexa Imaging is seeking an experienced AI Adoption Engineer, Clinical Imaging to lead the technical evaluation, validation, and successful clinical adoption of AI solutions across our imaging environment. Reporting to the SVP of AI Integrations and partnering closely with clinical leadership, the Clinical Applications team, and broader IT functions, this role owns the end-to-end process of bringing third-party clinical AI software from vendor selection through validated, deployment-ready recommendation.
This role sits at the intersection of clinical imaging workflow expertise, hands-on AI/ML implementation, and enterprise integration. The ideal candidate is equally comfortable installing and configuring vendor AI software across a range of deployment environments, designing validation methodologies that compare AI-generated results against radiologist findings using LLM/NLP techniques, and translating those findings into clear, defensible recommendations for live clinical deployment.
Success requires deep familiarity with radiology and imaging workflows from day one, including PACS, RIS, DICOM, HL7, and how radiologists read and report studies, combined with the ability to collaborate credibly with both clinical leadership and technical IT stakeholders.

Key Responsibilities
Clinical AI Deployment, Validation & Methodology

Install, configure, and maintain a diverse portfolio of third-party clinical AI software (including detection, classification, quantification, triage, decision support, and reporting solutions) across deployment environments in collaboration with Clinical Applications team, such as Lumexa's AI sandbox, vendor-provided edge servers, on-premise infrastructure, and cloud-based architectures – this role is expected to contribute to the decision-making of such best-fit architectures

Establish and operate DICOM routing, HL7 messaging, and de-identification workflows for both medical images and reports to support safe, compliant evaluation at scale
Design and execute structured validation studies tailored to each AI solution's clinical purpose, applying methodologies and metrics appropriate to the output type (e.g., sensitivity, specificity, discrepancy rates, clinical/workflow impact measures)
Build LLM/NLP-based comparison and analysis frameworks that systematically evaluate AI outputs against radiologist ground truth or other clinically relevant benchmarks
Build and continuously evolve Lumexa's clinical AI validation playbook as a flexible, reusable framework, including automated tooling that reduces time-to-evaluation through streamlined de-identification, cohort selection, and vendor data transfer
Author clear, evidence-based go/no-go recommendations for the AI Governance Council, including risk assessment and deployment scope

Clinical Leadership Partnership & Workflow Integration

Collaborate closely with the Chief Medical Officer, National Physician Leadership Board, and local clinical/operational leaders to validate clinical solutions, define ground truth, set acceptance thresholds, and ensure AI capabilities align with radiologist workflows and clinical priorities
Translate validated AI capabilities into deployment-ready integration designs that fit within radiologist reading workflows, and turn clinical feedback into actionable technical configurations to drive maximum fit
Map current-state vs. future-state workflows showing how AI outputs surface to radiologists, technologists, and operations teams
Identify workflow risks, change management considerations, and adoption barriers ahead of production deployment, and build trust with clinical leadership through clear communication and rigorous methodology
Build trust and credibility with clinical leadership through clear communication, rigorous methodology, and responsiveness to clinical input

Cross-Functional Technical Collaboration

Partner with the Clinical Applications team to ensure validated solutions are designed for production-portable deployment and to support the handoff from evaluation to live implementation in clinical technology stack (RIS, PACS, Reporting, etc.)
Engage Infrastructure to assess deployment architecture, edge server requirements, network considerations, compute and storage needs, and networking with existing environments
Engage Information Security to evaluate vendor security posture, data handling practices, encryption, access controls, and HIPAA compliance
Assess each vendor's overall technical maturity, including how the solution is built, supportability, scalability, and readiness for enterprise engagement, and surface risks early in the evaluation process

Vendor Technical Evaluation

Partner with the AI Integrations leadership team and Procurement during vendor selection by assessing technical fit, integration complexity, performance claims, and FDA/regulatory status
Conduct hands-on technical due diligence including reviewing model performance documentation, regulatory clearances, edge server architecture, and integration requirements
Support contract negotiations with technical input on SLAs, model retraining cadence, and performance guarantees

Continuous Improvement & Knowledge Leadership

Stay current on the rapidly evolving clinical imaging AI landscape, including new vendors, modalities, FDA clearances, and CPT reimbursement codes
Identify opportunities to expand Lumexa's clinical AI portfolio based on emerging capabilities
Contribute to thought leadership in AI governance and validation methodology

Required Qualifications

5+ years of experience in clinical imaging informatics, radiology AI deployment, or imaging AI vendor field engineeringCon
Hands-on experience installing and configuring clinical AI software across one or more deployment environments (sandbox, edge server, on-prem, or cloud), including end-to-end responsibility for data routing, anonymization, and system configuration
Demonstrated working knowledge of clinical imaging workflows, including how radiologists read studies, interpret findings and finalize reports
Strong fluency with imaging informatics standards: DICOM, HL7, FHIR, and PACS/RIS architecture
Hands-on technical skills with Python (or equivalent) for scripting validation pipelines, data extraction, and comparison analysis
Experience with LLM/NLP techniques for text comparison, semantic similarity, or structured information extraction from clinical reports
Demonstrated experience designing and executing AI performance validation studies, including defining metrics, ground truth, and study methodology
Demonstrated experience building or improving automated de-identification and data preparation pipelines for clinical AI vendor evaluations, including PACS cohort selection, DICOM header and burned-in pixel anonymization, paired report de-identification, and secure vendor packaging. Hands-on experience with DICOM routing and anonymization platforms (e.g., Laurel Bridge Compass, RSNA CTP) alongside complementary tooling (e.g., Presidio, AWS Comprehend Medical, OCR-based pixel masking) strongly desired
Proven ability to collaborate with both clinical leadership and technical IT stakeholders
Ability to operate independently in ambiguous, fast-moving environments with minimal oversight
Strong project scoping, prioritization, and execution skills across multiple concurrent vendor evaluations and projects

Preferred Qualifications

Basic clinical knowledge of radiology modalities and respective clinical workflows (e.g., CT, MRI, mammography, X-ray, ultrasound)
Experience at a clinical imaging AI vendor in a solutions, field, or implementation engineering capacity
CIIP (Certified Imaging Informatics Professional) certification or equivalent
Background as a radiology technologist, imaging informaticist, or radiology research engineer
Experience with cloud platforms (AWS, Azure, GCP) for AI workload deployment
Familiarity with FDA 510(k) clearance process and CPT reimbursement codes for imaging AI
Strong understanding of HIPAA and healthcare compliance requirements related to clinical AI
Advanced degree in biomedical engineering, medical imaging, computer science, or related field

Success Profile

Clinically fluent: Speaks the language of radiology and imaging workflows with credibility
Hands-on and pragmatic: Builds, installs, and validates directly rather than just designing on paper
Methodologically rigorous: Brings structure and reproducibility to validation, not ad-hoc testing
Cross-functionally credible: Earns the trust of IT, clinical and operational stakeholders
Self-starter: Identifies what needs to be done and drives it forward with minimal direction
Adaptable: Thrives in a fast-evolving AI landscape with new vendors, modalities, and capabilities arriving constantly
Continuous learner: Actively stays current on imaging AI capabilities, regulatory developments, and validation methodologies

Example Scope of Work – Modality and Vendor Example for Illustration Purpose Only

Installing a new chest CT AI vendor's edge server in the appropriate environment, configuring DICOM routing and de-identification, and running it against a curated set of exams to score AI-vs-radiologist concordance using LLM-based report comparison
Designing the validation study, ground truth methodology, and acceptance criteria for a new mammography AI module in partnership with lead breast radiologists and the CMO
Partnering with Infrastructure and Information Security to evaluate a vendor's edge server architecture, security posture, and overall technical maturity prior to procurement
Building the standard validation playbook applied to every new imaging AI module before live deployment
Monitoring post-deployment performance of live AI solutions and validating real-world results against pre-deployment assumptions and expectations, identifying drift, performance gaps, or unintended workflow impact, and recommending tuning, expansion, or retirement appropriate