Dev Docs

Technical explanation for engineers, learning scientists, and program designers implementing the Boyle System

Operational reference for institutional leads, IT administrators, and implementation teams deploying the Boyle System. Covers RAG vs. long-context tradeoffs, active pilot metrics, integration methods and stability, security data classifications by context, and target deployment scenarios. After reading, you can plan a technically sound deployment and assess data classification requirements for your institution.

Explanation document for research leads, institutional partners, and program evaluators assessing the Boyle System. Covers the reproducibility problem, NotebookLM RAG architecture, and the three-role AI partnership model. After reading, you can evaluate whether the Boyle System addresses your institution

Planning reference for institutional partners, program directors, and the Boyle System development team. Documents critical and high-priority action items with effort estimates, open architectural questions requiring partner input, and the full feature roadmap. After reading, partners can identify which open questions require their input and track the development trajectory of the system.

Full specification for the Medhavi 1.0 → 1.5 upgrade: what exists, what is being built, and why. Covers four AI modes, context awareness, signal collection, the Glimmer assignment, and build responsibilities.

Revised technical specification for the Medhavi intelligent textbook lexicon system. Defines two distinct pipelines: Wikipedia keyword detection for pop-up generation, and NanoLex fine-tuning for lexical entry extension. Written for Thejus Thomson, Prof. Brown, and Dhruv.

Explanation of the pedagogical architecture underlying the Medhavy adaptive learning platform. Covers the five learning approaches, the multi-armed bandit selection mechanism, why TEXTBOOK_ONLY is a pedagogical decision and not just a safety constraint, and what content must contain to support each approach. For developers and content leads who need to understand the system before building for it or extending it.

Technical specification, pedagogical framework, and business model for deploying the Medhavy adaptive learning platform under institutional branding. Covers multi-tenant architecture, five learning approaches, content pipeline, and white label implementation roadmap for business schools, think tanks, K-12 networks, and professional associations.

Technical reference for developers working on the Medhavy adaptive learning platform

Step-by-step deployment guide for developers onboarding a new institutional client to the Medhavy white label platform. Covers tenant registry setup, credential configuration, theme injection, persona calibration, Orama index provisioning, content pipeline handoff, and launch verification. After completing this guide, a developer can bring a new white label deployment from zero to live without assistance.

Field-level reference for researchers, fellows, and team leads using the Boyle System. Defines all six MVAL fields — What, Why, How, Environment, Results, Questions — with required content standards, sufficient vs. insufficient entry examples, and the failure artifact protocol. After reading, you can write a complete, compliant MVAL entry for any experiment or analytical session.

Operational reference for researchers and team leads managing NotebookLM corpora within the Boyle System. Covers source format performance, the Ouroboros note-to-source conversion technique, source stitching strategies, and notebook segmentation taxonomy. After reading, you can structure and maintain a NotebookLM corpus that maximizes retrieval quality and stays within platform limits.

Technical reference for the Boyle System — AI-assisted documentary infrastructure for scientific reproducibility. Covers MVAL protocol, corpus management, adaptive instructional design, and implementation roadmap for research labs, graduate programs, think tanks, and executive education.