Staff AI Solutions Engineer - Nuclear & Safety

Job ID: 2027Req ID: 353Location: Meriden, CT, USAPay: USD 180000 (Salary)Posted: 2026-10-21Pay Type: Salary As a global leader in radiation safety, detection and measurement, Mirion Technologies is advancing radiation safety through innovative technologies and solutions. Mirion's Nuclear & Safety business group builds the hardware, software, and services that keep nuclear power plants, laboratories and research facilities, defense operations, and radiation-hazard workplaces safe and effective. We're investing in a small, high-impact team of embedded AI Solutions Engineers who will sit inside our Nuclear & Safety business units and turn everyday operational friction across manufacturing, quality, service, and back-office workflows, into measurable efficiency gains using modern AI. This is not a research role and not a central-platform role. You will spend most of your time with the people who run the business (i.e. plant managers, quality engineers, field service leaders, product managers), finding the highest-leverage problems and then personally designing, building, and shipping the solution. Your toolkit spans the full range of applied AI: computer vision for visual inspection and image-based QA, spectral and sensor-data analytics, basic software development, and agentic AI / LLM-based automation for decision support and workflow orchestration. You'll report directly to the Chief AI and Digital Officer, giving you a direct line to enterprise AI strategy while you operate with the autonomy of a startup engineer inside a public, nuclear-regulated industrial company. What You'll Do Find and frame the opportunity Embed within the Nuclear & Safety business unit, attending their staff meetings, walking their production floor, shadowing their field service and quality teams to fully understand the business and landscape of problems Build a first-hand map of where manual, error-prone, or slow processes are costing time and money. Translate ambiguous, non-technical pain points ("our detector inspection and QA review takes too long," "we can't keep up with service tickets") into scoped, buildable AI solutions, and build the business case for prioritizing them. Build and ship Design and deploy solutions for use cases such as automated visual inspection of detectors and assemblies, defect and anomaly detection, spectral and sensor-data analysis, and image-based quality verification. Design and deploy agentic AI systems (LLM-based agents, workflow orchestration, and automation) to reduce manual effort in manufacturing, service, quality and compliance documentation, and other back-office functions. Own the full lifecycle: rapid prototyping, pilot with real users, hardening for production, and handoff or ongoing support, leveraging shared enterprise services and without waiting for a large team to do it for you. Communicate and drive adoption Present progress, trade-offs, and results in plain business language to business counterparts (plant managers, quality leaders, and executives who don't have an AI background) and adjust your approach based on their feedback. Define and track the metrics (cycle time, error rate, cost, throughput) that prove a solution is actually working, and course-correct quickly when it isn't. Lead change management and light training so solutions get adopted, not shelved. Operate within the broader AI function Partner with the Chief AI & Digital Officer and central AI/Digital team to stay aligned on architecture, security, data governance, and tooling standards, and to bring reusable patterns back to the center. Apply appropriate rigor for a regulated nuclear environment (nuclear quality assurance, validation, documentation, export controls, cybersecurity) in partnership with Quality, Regulatory, and Compliance colleagues. What Success Looks Like in Year One At least 2–3 AI-driven solutions in production, each tied to a measurable operational KPI (time saved, defect reduction, cost avoidance, throughput gained). A trusted, self-sustaining relationship with business unit leadership. They should come to you with problems rather than reaching out to the CAIDO, directly. A reusable playbook or pattern (for computer vision, for agentic workflows, or both) that can be extended to other Nuclear & Safety business units or replicated by peer AI solutions engineers. What You Bring Required 5+ years shipping production software or applied ML/AI systems, with real users depending on what you built. Experience in software engineering and cloud-based application development and deployment Hands-on experience building agentic or LLM-based systems (e.g., LangChain, LangGraph, AutoGen, or a comparable custom agent framework), including prompt design and retrieval-augmented approaches. Strong core software engineering: Python, cloud platforms (Azure, AWS, or GCP), APIs, and enough MLOps/DevOps fluency to get something into production and keep it running. A demonstrated track record of high agency. Ideally a history of projects you identified, scoped, and delivered with minimal direction, not just tickets you executed that were assigned in JIRA. Excellent communication skills with non-technical audiences. Strong ability to clearly explain what a model does and doesn't do without jargon, and knowledge of how to build trust with non-technical operators. Comfort with ambiguity and autonomy. You will often be the only AI/software person in the room and working daily with those with little-to-no AI skills. You will be part of a team of centralized peers with strong skills but will operate independently to deliver business-unit-specific results. Strongly Preferred Hands-on experience building computer vision solutions (e.g., OpenCV, PyTorch or TensorFlow, image classification/detection/segmentation pipelines), or ML on spectra, waveforms, and sensor time-series data. Experience in Azure or AWS, Databricks, and Claude/OpenAI integration via AWS Bedrock or Azure Foundry Experience in a regulated industry — nuclear power, defense, medical devices, or industrial manufacturing under a formal quality system (e.g., ISO 9001, 10 CFR 50 Appendix B, NQA-1). Prior experience as an embedded, business-facing, or "forward-deployed" engineer, solutions engineer, or applied AI generalist, rather than a purely research-facing ML role. Domain exposure to nuclear power plant instrumentation and controls, radiation detection and measurement, gamma spectroscopy, health physics, or radiation safety and monitoring workflows. Why This Role Direct visibility with the Chief AI & Digital Officer and a real hand in shaping how AI gets built and scaled across Mirion Nuclear & Safety. The rare combination of startup-style autonomy and ownership with the stability, resources, and mission of a global public company protecting workers, communities, and the environment. A builder's role. You will ship things that change how people work, not produce slide decks about how AI might someday help. Mirion Technologies (NYSE: MIR) is a global leader in radiation safety, science, and medicine, protecting people and the planet from the harmful effects of ionizing radiation. Mirion Nuclear & Safety powers Nuclear Power Plants, Laboratory & Research (including spectroscopy and scientific analysis), Radiation Safety & Monitoring, and Defense. Mirion offers a comprehensive benefits package that includes Vacation & PTO Time, medical, dental, and vision coverage, a 401(k) retirement savings plan with matching, HSA and FSA options, life and disability insurance, an Employee Assistance Program, wellness resources, and several voluntary benefits including legal assistance, identity protection, and pet insurance.