How bishwaghimire/ai-learning-roadmaps works
1. Repository type and purpose
- This is not an application codebase; it is a curated knowledge hub for AI/ML/DL learning.
- The architecture is centered on content organization, not runtime services or build pipelines.
2. Primary entry point
- README.md is the main navigation surface and effectively the “homepage.”
- It explains who the repo is for, how to use it, and the recommended learning sequence.
3. Core content subsystem: roadmap layer
- The roadmaps/ directory is the main structural backbone.
- Each markdown file is a standalone learning path for a domain:
- foundations: AI, Data Science, ML, Deep Learning
- specializations: CV, NLP, LLM, GenAI, RAG
- engineering/research tracks: MLOps, Research Scientist, AI Safety/Alignment
- These roadmaps are the main dependency graph for the repo: users typically move from broad foundations to specializations.
4. Reference library subsystem: books and PDFs
- The resources/ directory holds offline learning materials grouped by subject:
- foundations/
- Data-Science/
- ML/
- DL/
- AI/
- This is a curated document store, mainly PDFs, intended to support the roadmap items with deeper reading.
5. Presentation and branding assets
- assets/ai-image.png is the only visible media asset in the tree.
- It serves documentation/branding purposes inside the README, not product functionality.
6. Supplementary documentation and governance
- CONTRIBUTING.md defines the collaboration boundary for new content and curation rules.
- LICENSE sets the repository’s redistribution terms.
- Packages.md provides environment/tooling guidance for learners and acts as a setup reference.
7. External knowledge and service boundaries
- The repo heavily depends on external links for videos, courses, communities, papers, and cloud/platform docs.
- These include YouTube, MIT OCW, Coursera, arXiv, Hugging Face, OpenAI/Google/DeepMind blogs, and community platforms like Reddit, Discord, Telegram, and Stack Overflow.
- Architecture-wise, the repo is a curated index that routes users to external learning systems rather than hosting them.
8. How information flows through the repo
- A learner starts in README.md.
- From there, they choose a roadmap in roadmaps/.
- Each roadmap points to relevant concepts and often to resources in resources/ or external references.
- The flow is essentially: overview → structured path → deep reference material → external practice/learning sources.
9. Content layering
- The repo separates content into distinct levels:
- orientation and sequencing in README.md
- detailed progression in roadmaps/
- long-form source material in resources/
- environment/tooling guidance in Packages.md
- This layered structure is important for diagramming because it shows a clear top-down learning funnel.
10. Domain boundaries
- The repository spans multiple AI subdomains but keeps them logically isolated.
- For example, ML/DL foundations are separated from RAG, LLMs, MLOps, and AI safety, while still linked through the roadmap hierarchy.
- This prevents the repo from becoming a flat list of links and gives it a modular architecture.
11. Technology relevance
- The repo is documentation-first, but the tooling it references shapes the ecosystem:
- Python, Jupyter/Colab for learning
- PyTorch, TensorFlow, JAX for modeling
- Hugging Face, LangChain, LlamaIndex, vector DBs, MLflow/W&B/DVC for modern AI workflows
- These are not implemented in the repo, but they define the expected downstream stack for learners.
12. What is intentionally absent
- There is no code execution layer, no CI/CD, no backend, no database, and no packaged app.
- So the architecture diagram should emphasize curation and navigation, not runtime components.
13. Best diagram framing
- Model the repo as a content platform with three main zones:
- navigation/index (README.md)
- structured learning paths (roadmaps/)
- offline reference materials (resources/)
- Add external services as outbound dependencies, since much of the educational value comes from links out to third-party content.
Main entry points: roadmaps Roadmaps · resources Resource library · README.md README · ml-roadmap.md Foundations
Components 32
Navigation
Top-level orientation, usage, and governance for the knowledge hub.
- README · Main homepage and navigation surface for learners.
README.md - Contributing · Rules and boundary for adding or curating content.
CONTRIBUTING.md - Packages · Environment and tooling reference for learners.
Packages.md - License · Repository redistribution terms.
LICENSE - Brand asset · Documentation image used in the README.
Roadmaps
Structured learning paths across AI/ML subdomains.
- Roadmaps · Collection of domain-specific learning sequences.
roadmaps - Foundations · Core baseline for AI, data science, ML, and deep learning.
roadmaps/ml-roadmap.md - AI · Broad artificial intelligence orientation and concepts.
roadmaps/ai-roadmap.md - Data Science · Data analysis, statistics, and practical DS progression.
roadmaps/data-science-roadmap.md - ML · Machine learning methods, theory, and practice.
roadmaps/ml-roadmap.md - Deep Learning · Neural network foundations and modern DL workflow.
roadmaps/deep-learning-roadmap.md - Specializations · Focused tracks built on the core foundations.
- Computer Vision · Visual perception and image understanding track.
roadmaps/computer-vision-roadmap.md - NLP · Language understanding and text processing track.
roadmaps/nlp-roadmap.md - LLM · Large language model concepts and ecosystems.
roadmaps/llm-roadmap.md - GenAI · Generative AI workflows and applications.
roadmaps/generative-ai-roadmap.md - RAG · Retrieval-augmented generation and knowledge grounding.
roadmaps/rag-roadmap.md - Ops & Research · Production, research, and safety-oriented tracks.
- MLOps · Production AI, deployment, and lifecycle management.
roadmaps/mlops-production-ai-roadmap.md - Research Scientist · Research workflow, reading, and experimentation track.
roadmaps/research-scientist-roadmap.md - AI Safety · Alignment, safety, and responsible AI track.
roadmaps/ai-safety-alignment-roadmap.md
Resources
Offline reference library of books and PDFs.
- Resource library · Curated offline books and PDFs supporting the learning paths.
resources - Foundations books · Python and deep learning primers for the baseline layer.
resources/foundations - Data Science books · Math, statistics, SQL, and interview preparation materials.
resources/Data-Science - ML books · Machine learning theory and implementation references.
resources/ML - DL books · Deep learning reference material.
resources/DL - AI books · General AI reference material.
resources/AI
External sources
Outbound learning systems and communities linked from the repo.
- External links · Third-party courses, papers, blogs, tools, and communities referenced by the repo.
- Courses · MOOCs and lecture series such as MIT OCW and Coursera.
- Papers · Research sources such as arXiv and related technical writing.
- Platforms · Tool and ecosystem docs such as Hugging Face and major AI vendor blogs.
- Communities · Discussion and support channels such as Reddit, Discord, Telegram, and Stack Overflow.
Connections 34
- README → Roadmaps: starts here · Learners begin with the repo homepage and choose a path.
- README → Resource library: references · The homepage also points to curated books and PDFs.
- README → Packages: setup · Learners can consult environment and tooling guidance.
- README → Contributing: rules · Content additions and curation follow the contribution guide.
- README → External links: outbound · Navigation often routes users to outside learning systems.
- Roadmaps → Foundations: baseline · The foundational layer underpins the rest of the learning paths.
- Roadmaps → AI: domain · Broad AI orientation sits alongside the core baseline.
- Roadmaps → Data Science: domain · Data science is one of the primary entry paths.
- Roadmaps → ML: domain · Machine learning is a central track in the set.
- Roadmaps → Deep Learning: domain · Deep learning extends the core foundation layer.
- Roadmaps → Specializations: branches · Specialized tracks branch off from the common foundations.
- Specializations → Computer Vision: track · Computer vision is one specialization branch.
- Specializations → NLP: track · NLP is one specialization branch.
- Specializations → LLM: track · LLM learning is one specialization branch.
- Specializations → GenAI: track · Generative AI is one specialization branch.
- Specializations → RAG: track · RAG is one specialization branch.
- Roadmaps → Ops & Research: advanced paths · Production, research, and safety are separate advanced branches.
- Ops & Research → MLOps: track · MLOps covers production AI workflows.
- Ops & Research → Research Scientist: track · Research scientist guidance supports experimentation and reading.
- Ops & Research → AI Safety: track · AI safety and alignment form a distinct advanced path.
- Foundations → Foundations books: supported by · Baseline roadmap items are reinforced by foundational books.
- Data Science → Data Science books: supported by · Data science topics are backed by curated references.
- ML → ML books: supported by · Machine learning concepts link to deeper reading materials.
- Deep Learning → DL books: supported by · Deep learning concepts link to deep reference material.
- AI → AI books: supported by · General AI study is reinforced by reference texts.
- Resource library → Foundations books: contains · The resource library includes foundational materials.
- Resource library → Data Science books: contains · The resource library includes data science materials.
- Resource library → ML books: contains · The resource library includes machine learning materials.
- Resource library → DL books: contains · The resource library includes deep learning materials.
- Resource library → AI books: contains · The resource library includes general AI materials.
- External links → Courses: includes · Courseware and lecture links are part of the external index.
- External links → Papers: includes · Research papers and preprints are part of the external index.
- External links → Platforms: includes · Tooling and vendor ecosystems are part of the external index.
- External links → Communities: includes · Community and support channels are part of the external index.