This repository offers a comprehensive, practical roadmap for mastering Machine Learning, Deep Learning, LLMs, Generative AI, and MLOps, targeting aspiring ML engineers, data scientists, and researchers. It emphasizes hands-on project-based learning, real-world scenarios, and efficient learning strategies, guiding users from foundational concepts to production-level skills with a focus on portfolio development.

How It Works

The roadmap is structured into modular tracks covering Python, classical ML, deep learning, LLMs, Generative AI, MLOps, and specialization, integrating numerous practical exercises and curated free resources. It promotes a "vibe coding" methodology, leveraging LLMs as coding partners, and emphasizes building a portfolio through tangible artifacts like code repositories and deployed demos.

Quick Start & Requirements

• Installation: Requires Python 3.12+, VS Code/Cursor, Git, a GitHub account, and a Kaggle account.
• Prerequisites: Standard development environment; no specific hardware is mandated beyond typical development needs, though GPUs are recommended for deep learning experimentation.
• Time Commitment: Significant, estimated at 9-18 months for junior-level proficiency with 10-15 hours/week.
• Resources: Extensive links to free courses (Coursera, fast.ai, Hugging Face, etc.), books, Telegram channels, and communities are provided.

Highlighted Details

• Project-Based Learning: Features 60 practical coding tasks and multiple capstone projects designed to build a demonstrable portfolio.
• "Vibe Coding" Methodology: Integrates LLMs (Claude, GPT) as coding partners for tasks from refactoring to end-to-end feature development.
• Modular Course Structure: Offers distinct, self-contained courses within the repository covering foundational math, neural networks, data science, LLM engineering, CV engineering, and agent development.
• Comprehensive Career Guidance: Includes detailed career tracks, skill level definitions (Junior to Guru), interview preparation strategies, and market insights.
• Curated Resource Lists: Provides extensive, regularly updated lists of free learning materials, papers, and community channels.

Maintenance & Community

The project welcomes contributions via Pull Requests for updates and improvements. It actively references and links to numerous Russian and English-speaking ML/AI communities, including Telegram channels, ODS.ai Slack, Hugging Face Discord, and EleutherAI Discord.

Licensing & Compatibility

Released under the MIT License, permitting free use, forking, and adaptation for both personal and commercial projects.

Limitations & Caveats

The roadmap is extremely comprehensive, demanding a substantial time commitment (9-18 months minimum). It relies heavily on self-directed learning. While cloud options are suggested, local deep learning experimentation may require dedicated GPU hardware. The "vibe coding" approach may need integration adjustments depending on existing team workflows.