PhD Intern - Computer Science

Pacific Northwest National Laboratory

1d•$24 - $36•Onsite

About The Position

The Electricity Infrastructure and Buildings Division of PNNL is accelerating the transition to a decarbonized, efficient, resilient, and secure energy system through basic and applied research. We leverage a strong technical foundation in power and energy systems and advanced data analytics to drive innovation, transform markets, and shape energy policy. The Optimization and Control Group and the Energy Systems and Resilience Group are looking for a Summer Intern to contribute to innovative research on foundational models and cybersecurity. The intern will play a key role in leveraging foundational models to streamline user interactions and provide targeted insights for an existing multi-criteria decision-making platform used in large-scale infrastructure planning. Additionally, the position will also focus on the application of Artificial Intelligence (AI) to develop a set of supplementary tools to enhance the Cybersecurity Capability Maturity Model (C2M2). The intern will work with PNNL staff to design, implement, and test an AI agent for two specific tasks: (1) simplify user interactions and answer complex infrastructure planning queries for multi-criteria decision making in large-scale infrastructure planning, and (2) enable users of the C2M2 to advance the maturity of their cybersecurity program. The work will require an understanding of AI training, testing, and validation. Furthermore, the intern may be required to work with PNNL staff in the deployment of the tool.

Requirements

Currently enrolled/matriculated in a PhD program at an accredited college.
Minimum GPA of 3.0 is required.
Intermediate programming ability in Python; familiarity with software engineering basics (Git, code reviews, unit testing, packaging).
Some working knowledge of LLMs and agent frameworks (e.g., OpenAI/Anthropic APIs, Azure OpenAI; LangChain and/or LlamaIndex) including prompt design, tool/function calling, retrieval-augmented generation (RAG), and basic agent orchestration.
Experience with data preparation and knowledge management: cleaning/structuring text, document ingestion, metadata/tagging, embeddings, and working with vector databases (e.g., FAISS, Chroma, Pinecone, Azure AI Search, Elasticsearch).
Understanding of AI evaluation, testing, and validation: creating test sets, measuring response quality/accuracy, hallucination and bias considerations, regression testing for prompts/agents, and documenting limitations.
Familiarity with cybersecurity concepts sufficient to work with C2M2 content (e.g., governance, risk management, asset management, incident response); ability to learn the C2M2 model quickly and translate it into user workflows.
Basic experience or understanding building web apps or APIs for deployment/integration (e.g., Flask/FastAPI, Streamlit, or similar), plus containerization fundamentals (Docker) and cloud/service deployment awareness.
Strong written communication and documentation skills (requirements capture, design notes, user guidance), ability to collaborate with researchers/engineers, and comfort working with sensitive information under defined controls.

Nice To Haves

Currently enrolled in PhD program Computer Science, Electrical Engineering, Computer Engineering, Data Science, Cybersecurity, or a closely related field.
Preferred coursework or project experience in: machine learning/NLP, information retrieval, software engineering, and/or cybersecurity.
Relevant certifications or training such as Azure/AWS fundamentals, Security+ (or equivalent), or prior research/industry experience with applied AI.

Responsibilities

Design, implement, and test an AI agent for two specific tasks: (1) simplify user interactions and answer complex infrastructure planning queries for multi-criteria decision making in large-scale infrastructure planning, and (2) enable users of the C2M2 to advance the maturity of their cybersecurity program.
Work with PNNL staff in the deployment of the tool.