Intern - AI Engineer

AI Engineer Intern

Nelnet's AI Lab is building the agent frameworks, integrations, and technical infrastructure that will define how Nelnet operates with AI over the next several years. AI Engineer Interns join a small, high-velocity team doing that work directly: building and testing agents, exploring LLM-based workflows, researching emerging frameworks, and contributing to production-bound tooling.

Agent Development:
Help build, test, and iterate on AI agents using the team's secure agent framework. This includes working with LLM APIs, multi-agent orchestration patterns, and tooling that runs in Nelnet's Microsoft 365 and cloud environment.

Technical Research:
Track developments in LLM frameworks, agent architectures, tooling, and industry approaches. Evaluate options against the team's real constraints (security, identity, M365 integration) and surface findings with clear "so what" framing. Research here feeds real decisions, not slide decks.

Integration and Prototyping:
Build proof-of-concept integrations between LLM-based systems and Nelnet's internal tools and APIs. Test assumptions early, document what worked and what didn't, and hand off something the team can build on.

Code Review and Documentation:
Document design decisions, system behavior, and implementation notes with the precision of someone who expects a future engineer to read it. Participate in code review as both author and reviewer.

Final Presentation:
Present your work, findings, and recommendations to AI Lab leadership and team members at the end of the internship period.

Education

Pursuing a degree in Computer Science, Software Engineering, Data Science, or a related technical field.

We care more about what you've built than which courses you've taken. A strong portfolio of personal projects, open-source contributions, or prior technical internships carries more weight than GPA alone.

Experience

Hands-On AI Tool Usage (Required):
Has used LLMs for real technical work, not just experimentation. This might mean using Claude or ChatGPT to write, debug, or extend code; building a personal project with an LLM API; or integrating AI tooling into a workflow. Can speak concretely about what they built, how it worked, and what the limits were.

Programming:
Python is the primary language. Solid working knowledge is required. Experience with any of the following is a strong plus:
Lang Chain, Lang Graph, Claude API, OpenAI/Claude SDK, FastAPI, or similar frameworks. Familiarity with REST APIs, JSON, and basic cloud or serverless patterns is beneficial.

Prior Technical Projects:
Personal projects, research, coursework projects, or prior internships that involved building something end-to-end. We want to see evidence of independently scoped and completed technical work, however small.

Competencies

LLM Fluency:
Uses LLMs as a genuine engineering tool, not just a novelty. Understands prompt construction, context windows, tool use, and the failure modes that matter in production. Knows when to reach for an LLM and when not to.

Agentic Systems Thinking:
Understands how business workflows can be decomposed into agentic patterns: what an agent owns, what it delegates, what triggers it, and where it can fail. Asks the right questions about data access, identity, security, and trust before assuming the happy path.

Programming and Technical Depth:
Working Python proficiency. Comfortable reading unfamiliar code, debugging across system boundaries, and writing code others can maintain. Experience with LLM frameworks or API integration is a significant plus.

Research Rigor:
Can evaluate a new framework or approach systematically: what problem it solves, what the tradeoffs are, and whether it fits this team's constraints. Not satisfied by marketing copy or surface-level comparisons. Produces findings that are actionable.

Problem-Solving Under Ambiguity:
Early-stage environments don't have complete specs. Comfortable breaking down a vague requirement into concrete next steps, identifying what needs to be true before moving forward, and flagging when a direction isn't working.

Communication:
Can explain a technical approach to a non-technical stakeholder and a design decision to a senior engineer in the same day. Written communication is precise. Code and documentation reflect the same clarity as verbal explanations.

Curiosity Over Credentials:
More interested in understanding how something actually works than in appearing to already know it. Experiments readily, is not defensive about being wrong, and learns faster from real usage than from documentation alone.

Ethical Grounding:
Understands that AI systems reflect design choices and that those choices have real consequences. Thinks seriously about data privacy, identity, and the responsible use of AI in an enterprise setting. Can engage with the tradeoffs, not just recite the principles.

Adaptability:
The AI tooling landscape changes fast and so does this team's priorities. Comfortable updating assumptions when new information…