Unpaid AI Engineering Intern (Search & Retrieval Systems)

About the Role

Dynamic Active is looking for an AI Search & Retrieval Systems Intern to lead the development of our historical lookup and analytical feature set within our teacher evaluation and classroom observation software. In this role, you will help design and implement the systems that allow principals, teachers, and district evaluators to instantly find, analyze, and surface patterns across past classroom evaluations.

While our transcription engine and core grading logic are managed by other teams, this internship is focused specifically on the data retrieval and historical synthesis layer. You will tackle the challenge of making large volumes of past observation data highly accessible and actionable—evaluating when to use traditional database indexing versus when to implement cost-effective AI solutions like Semantic Search (Embeddings) and Retrieval-Augmented Generation (RAG).

You’ll work alongside our development team to investigate vector search approaches, manage context windows efficiently to minimize token costs, and build prototypes that bridge traditional data filtering with intelligent semantic discovery. This is a hands-on opportunity to solve real product infrastructure challenges in an active EdTech environment.

The internship is mentorship-driven and designed around applied learning, technical execution, and meaningful product contributions.

What You’ll Work On

• Semantic Search Architecture: Research and prototype semantic search pipelines using text embedding models to let users query past evaluations with natural language (e.g., searching for concepts like "classroom pacing" or "student engagement strategies" instead of just names or dates).
• Cost-Efficient RAG Pipelines: Explore and build Retrieval-Augmented Generation architectures to synthesize historical data securely, ensuring Gemini or other foundation models are used efficiently to manage token usage and control API costs.
• Hybrid Search Implementation: Work with our engineering team to integrate traditional database filtering (SQL/NoSQL) with AI-powered search, giving users 100% accurate structural filters along with smart conceptual matching.
• Longitudinal Trend Analytics: Build proof-of-concepts that utilize LLMs to analyze historical data trends, helping administrators track a teacher’s progress or spot school-wide coaching opportunities across multiple past records.
• Cost & Token Benchmarking: Set up tracking and benchmarking to strictly evaluate the financial feasibility of AI features, analyzing the exact token costs per query to help leadership make data-driven architecture decisions.

Technologies & Areas of Interest

Experience with some of the following is helpful:

• Python
• Vector Databases / Libraries (e.g., Pinecone, Chroma, pgvector, or FAISS)
• Text Embeddings & Semantic Search (Google Vertex AI Embeddings, OpenAI, or Hugging Face)
• LLM Orchestration Frameworks (e.g., LangChain, LlamaIndex, or native Python SDKs)
• Relational Databases (SQL) and indexing structures
• Basic knowledge of cloud data pipelines and data privacy standards (highly relevant for educational data)

You do not need experience with every technology listed. We are interested in candidates who have a strong technical foundation and the desire to master these architectures.

Who We’re Looking For

This opportunity is designed for students, recent graduates, or early-career engineers who want practical experience working on data-dense AI workflows, search engineering, and cost-optimized system architecture.

The ideal candidate is:

• Intensely curious about information retrieval, search ranking, and database performance.
• Comfortable programming, preferably in Python, with a grasp of backend data flows.
• Interested in the math and mechanics behind semantic search and vector spaces.
• Cost-conscious—enjoys figuring out how to achieve smart results with minimal token usage.
• Excited to take a feature from the conceptual stage through active prototyping in an early EdTech startup.

What You’ll Gain

• Deep, practical knowledge of building production-level Semantic Search and RAG systems.
• Hands-on experience optimizing and benchmarking API token consumption for a commercial software product.
• Direct exposure to hybrid software architectures that pair traditional databases with modern generative AI.
• Professional networking, mentorship, and a clear understanding of startup product priorities.
• A formal letter of recommendation upon completion of the internship.
• Potential consideration for future paid opportunities based on performance and company growth.

California Legal Compliance

This is an unpaid volunteer internship intended to provide educational and professional-development experience.

The internship is structured around the intern as the primary beneficiary, with an emphasis on learning, mentorship, training, experimentation, and professional development. Intern activities are intended to complement rather than replace the work of paid employees.

Applicable legal protections, including protections against discrimination and harassment, continue to apply as required by law.

Internship Details

• Position: AI Search & Retrieval Systems Intern
• Type: Unpaid Internship / Volunteer
• Focus: Semantic Search, Embeddings, RAG Pipelines, Hybrid Databases, Cost Optimization
• Industry: Education Technology (EdTech)

Why Dynamic Active?

You’ll have the opportunity to build a high-visibility feature designed for real-world school administrative workflows. Instead of writing simple CRUD apps, you will be exploring the balance between database performance, user intuition, and AI API costs. If you want to dive deep into how software smartly handles historical records and trends, we encourage you to apply.