AI Engineer - Mission Innovation Lab

About the role

The SEI AI Division focuses on applied artificial intelligence and the engineering of AI technologies and systems. We lead a community-wide movement to mature AI Engineering for Defense and National Security, working with government customers to adopt AI and machine learning for mission capabilities.

Responsibilities

• Design, develop, and fine-tune a variety of AI models.
• Design autonomous agents and multi-step pipelines using LangChain, ReAct, tool-calling, or custom orchestration; employ the Model Context protocol to manage stateful interactions.
• Create Retrieval-Augmented Generation pipelines that combine external knowledge bases with LLMs to improve factual accuracy for warfighting applications.
• Implement end-to-end data pipelines, ETL processes, and back-end services (Python, C/C++, Java) that feed data to models.
• Build CI/CD pipelines for model training, validation, containerized deployment (Docker/Kubernetes), and security scanning; maintain model registries, monitoring, and version control of context protocols.
• Produce rapid prototypes, run benchmarks, and conduct robustness/adversarial testing in realistic environments.
• Work closely with senior ML engineers, software developers, and government customers; mentor junior staff and contribute to design reviews and documentation.
• Stay current with emerging LLM architectures, agentic paradigms, PEFT/LoRA methods, and AI-safety techniques; translate new research into operational capabilities.

Requirements

• Bachelor’s degree in Computer Science, Machine Learning, Statistics, Applied Mathematics, or a related field with at least eight (8) years of relevant experience, or a MS degree in the same with at least five (5) years of relevant experience.
• You will be subject to a background investigation and must be able to obtain and maintain an active Department of War (DoW) security clearance.
• You must be able and willing to work onsite 5 days per week at an SEI office in either Pittsburgh, PA or Arlington, VA.
• Proficiency in Python and at least one compiled language (C/C++ or Java); experience with REST/GraphQL APIs and containerization.
• Strong grasp of ML theory (supervised, unsupervised, reinforcement learning) and evaluation metrics.
• Hands-on experience fine-tuning LLMs and using frameworks such as Hugging Face Transformers, LangChain, or comparable agent tools.
• Familiarity with building RAG pipelines (vector stores, dense/sparse retrievers).
• Experience applying PEFT/LoRA methods (e.g., LoRA, adapters) to large models.
• Understanding of Model Context protocols for managing model state across multi-turn interactions.
• Experience building evaluation frameworks, benchmarks, or data quality pipelines.
• Experience with TensorFlow, PyTorch, or JAX; knowledge of data-pipeline tools (Airflow, Prefect, Ray) is a plus.
• Awareness of DevSecOps practices (CI/CD, GitOps, container security scanning, model-registry concepts) is desirable.

Qualifications

• Bachelor’s degree in Computer Science, Machine Learning, Statistics, Applied Mathematics, or a related field with at least eight (8) years of relevant experience, or a MS degree in the same with at least five (5) years of relevant experience.
• You will be subject to a background investigation and must be able to obtain and maintain an active Department of War (DoW) security clearance.
• You must be able and willing to work onsite 5 days per week at an SEI office in either Pittsburgh, PA or Arlington, VA.
• Proficiency in Python and at least one compiled language (C/C++ or Java); experience with REST/GraphQL APIs and containerization.
• Strong grasp of ML theory (supervised, unsupervised, reinforcement learning) and evaluation metrics.
• Hands-on experience fine-tuning LLMs and using frameworks such as Hugging Face Transformers, LangChain, or comparable agent tools.
• Familiarity with building RAG pipelines (vector stores, dense/sparse retrievers).
• Experience applying PEFT/LoRA methods (e.g., LoRA, adapters) to large models.
• Understanding of Model Context protocols for managing model state across multi-turn interactions.
• Experience building evaluation frameworks, benchmarks, or data quality pipelines.
• Experience with TensorFlow, PyTorch, or JAX; knowledge of data-pipeline tools (Airflow, Prefect, Ray) is a plus.
• Awareness of DevSecOps practices (CI/CD, GitOps, container security scanning, model-registry concepts) is desirable.