BEATRICE

BEATRICE (Briefing, Evaluation, Assistance, Tutoring, Reasoning, and Instructional Curriculum Engine) is the AI Teaching Assistant for Inquiry Institute. She serves as the primary academic support system for pupils—guiding them through coursework, answering questions, evaluating assignments, and helping them reason through complex problems in AI, engineering, and related fields.

Unlike a simple chatbot, BEATRICE functions as an active pedagogical agent embedded in the student’s working environment. She helps students interpret assignments, debug code, explore concepts, and develop their own solutions. Rather than simply providing answers, BEATRICE frequently uses a Socratic tutoring approach, encouraging learners to articulate their reasoning and iterate on their work. When deeper expertise is required, she can route questions to specialized AI faculty personas or reference curated course materials through retrieval systems.

Primary Interface

The primary interface for BEATRICE is a Chrome Extension, allowing her to operate directly within the student’s everyday workflow. This approach enables BEATRICE to assist students wherever they are working—whether reading course materials, browsing documentation, writing code, or interacting with online resources.

Through the browser extension, BEATRICE can:

• Provide contextual explanations of material students are reading
• Assist with coding tasks in GitHub or browser-based IDEs
• Summarize research papers and documentation
• Guide students through course exercises
• Capture and organize notes for later study
• Interact with the student’s Personal AI environment

This design makes BEATRICE feel less like a separate application and more like an ambient academic assistant that accompanies the learner across the web.

Technology Stack

BEATRICE is built on a modular, open infrastructure designed to support both cloud-based and sovereign local AI deployments.

Interface Layer

• Chrome Extension (primary UI)
• Web-based PWA interfaces for mobile and desktop access
• Integration with GitHub and browser-based development environments

Agent Orchestration

• LangChain / LangGraph style agent frameworks
• Tool-based architecture allowing BEATRICE to invoke specialized capabilities (RAG, evaluation tools, code execution, etc.)
• Task specifications and structured prompts to guide reasoning workflows

Knowledge Retrieval (RAG)

• Course materials, lecture notes, and textbooks indexed in vector databases
• Supabase + pgvector or similar vector search systems
• Retrieval pipelines that ground responses in verified course content

Model Layer

• Open-weight LLMs such as Qwen-Coder, Llama, and Mistral
• Cloud-hosted inference (Google Cloud Run, AWS Bedrock, or OpenRouter)
• Optional local inference for sovereign deployments (Mac Mini, Raspberry Pi clusters, or edge devices)

Development Environment Integration

• GitHub repositories as canonical sources for assignments and course materials
• GitHub Codespaces for student development environments
• Repository-based evaluation pipelines for grading and feedback

Evaluation and Feedback

• Automated testing of student code
• Assignment validation scripts
• Structured feedback and progress tracking tied to the student’s repository

Role in the Inquiry Ecosystem

Within the Inquiry ecosystem, BEATRICE acts as a bridge between students, faculty AI personas, and course infrastructure. She ensures that every student has access to consistent guidance, structured feedback, and personalized support regardless of class size.

Because BEATRICE operates through the browser and integrates with repository-based workflows, she transforms the learning process into a traceable and collaborative knowledge-building process. Prompts, revisions, reasoning steps, and solutions become part of the student’s academic record—allowing both instructors and learners to reflect on how understanding evolves over time.

In essence, BEATRICE is designed to be a persistent AI teaching assistant that accompanies the learner everywhere they study, combining modern AI models, retrieval systems, and browser-native tools to create a seamless AI-supported educational experience.