About the role

We are building ML systems that accelerate nuclear regulatory workstreams and augment reactor engineering and operations. You’ll design and deploy language- and simulation-driven models that automate documentation, support human-in-the-loop decision making, and inform optimization of plant and control strategies. The ideal candidate blends deep hands-on ML experience with rigor suitable for high‑assurance environments.

What you’ll do

Model & application development

• Design and implement large language model-driven solutions for nuclear regulatory, engineering, and operational contexts, within an agentic framework.
• Develop RAG and tool‑use flows over internal and public corpora (e.g., regulatory guidance, procedures, technical reports) to ground outputs and reduce hallucinations.
• Build evaluation harnesses and guardrails for reliability, transparency, and traceability of model outputs.

Safety & compliance automation

• Create pipelines that automate regulatory documentation generation, review, and compliance mapping.
• Implement test plans and acceptance criteria aligned with high‑assurance software practices (versioning, reproducibility, traceability).
• Package evidence (tests, evals, datasets, provenance) to demonstrate robustness and fitness for use in regulated workflows.

Reactor systems & optimization

• Apply reinforcement learning, simulation, and optimization (e.g., Bayesian methods, evolutionary algorithms) to improve design and operational parameters.
• Collaborate with nuclear engineers to translate physics and safety models into ML‑driven decision‑support tools and human‑in‑the‑loop workflows.
• Build ML systems that support autonomous and semi‑autonomous reactor control and monitoring in collaboration with controls & safety teams.

Platform & MLOps

• Establish scalable infrastructure for training, testing, and deployment (containerized services, experiment tracking, CI for ML, dataset/version governance).
• Design for secure, possibly air‑gapped and on‑prem environments; integrate with monitoring/observability for models in production.
• Interact and coordinate with external technology platform and solution providers.

Cross‑functional collaboration

• Work with engineering, regulatory affairs, product, and operations to align ML capabilities with nuclear and data‑center applications.

Qualifications

Required

• Strong programming in Python and modern ML frameworks (PyTorch, TensorFlow, or JAX), plus experience with Hugging Face or similar tooling.
• Demonstrated experience fine‑tuning and evaluating foundation models (prompting, adapters/LoRA, distillation, safety/robustness evals).
• Proven ability to build end‑to‑end ML applications: data pipelines, retrieval/grounding, inference services, and monitoring.
• Familiarity with safety‑critical software practices: version control, testing, reproducible ML workflows, change control.
• Comfortable working fully on‑site in Austin and collaborating closely with domain experts across disciplines.

Preferred

• Background in numerical methods, scientific computing, or physics‑based modeling.
• Experience with optimization methods (Bayesian optimization, evolutionary algorithms, reinforcement learning) for complex engineering systems.
• Exposure to high‑assurance or functional‑safety environments (e.g., IEC 61508 or analogous) and quality systems (e.g., NQA‑1‑style rigor) is a plus.
• Practical MLOps experience with distributed training/inference (e.g., Kubernetes/Ray), experiment tracking, and data/version governance.
• Familiarity with nuclear systems, energy infrastructure, or regulatory processes.

Physical

• Primarily office/lab work on‑site in Austin; occasional hands‑on work with compute and test equipment; ability to use a computer for extended periods.