Prompt Engineering Intern

About the Role

Agentic Academy Labs is looking for a Prompt Engineering Intern to join our Engineering team and help shape how students learn to build with LLMs. This is a hands-on, remote internship where you'll design, test, and iterate on prompts that power our curriculum, evaluation systems, and internal AI tooling.

You'll work directly with the founding engineering team to build eval harnesses, refine agent workflows, and ensure content quality across our learning platform. If you're fascinated by how prompt structure changes model behavior, love systematic experimentation, and want to see your work impact thousands of learners — this is the role for you.

What You Will Do

• Design, version, and maintain prompts for curriculum generation, code explanation, debugging assistance, and assessment rubrics
• Build and extend evaluation harnesses (automated + human-in-the-loop) to measure prompt quality, consistency, and pedagogical effectiveness
• Collaborate on agent workflows that chain LLMs, tools, and retrieval for multi-step educational tasks
• Run controlled experiments (A/B, prompt ablations, few-shot vs. zero-shot) and document findings in shared playbooks
• Partner with curriculum engineers to translate learning objectives into reliable prompt templates
• Contribute to our internal prompt library — a version-controlled, tested collection of reusable prompt components
• Write clear documentation and decision logs so the team (and future interns) can build on your work

Why Join Us

• Real impact: Your prompts ship to production and directly affect how students learn AI development
• Mentorship-first: Pair with senior engineers who've built LLM products at scale; weekly 1:1s and code/prompt reviews
• Open experimentation: Dedicated compute budget for running evals, trying new models, and publishing learnings
• Portfolio gold: Ship a public prompt engineering case study (with metrics) by the end of your internship
• Remote-first culture: Async-friendly, timezone-aware, with optional sync sessions for deep collaboration
• Learning community: Access to all Agentic Academy courses, office hours with industry practitioners, and a peer cohort of interns across disciplines

Required

• Strong grasp of LLM behavior — you've spent serious time prompting GPT-4, Claude, or open models and can explain why a prompt works (or fails)
• Comfortable with Python (or TypeScript) for building eval scripts, running batch experiments, and integrating with our tooling
• Experience designing structured prompts: system messages, few-shot examples, chain-of-thought, function calling schemas
• Familiarity with evaluation concepts: golden sets, LLM-as-judge, semantic similarity, human annotation workflows
• Ability to write clean, readable markdown documentation and communicate experiment results clearly
• Self-directed learner who treats prompt engineering as an empirical discipline, not guesswork

Learning Goals (What You'll Gain)

• Production-grade prompt lifecycle: versioning, testing, deployment, monitoring
• Building eval harnesses that catch regressions before they hit learners
• Agent orchestration patterns (ReAct, plan-and-execute, multi-agent debate)
• Collaborating with curriculum designers to ground prompts in pedagogical research
• Open-source contribution practices — we upstream useful patterns

Nice to Have

• Prior internship or project work with LangChain, LangGraph, DSPy, Instructor, or similar frameworks
• Experience with RAG evaluation (retrieval quality, citation faithfulness, answer relevance)
• Background in CS education, technical writing, or instructional design
• Published blog post, notebook, or repo demonstrating systematic prompt experimentation
• Familiarity with Weights & Biases, MLflow, PromptLayer, or similar observability tools