iturdikulov.com logoiturdikulov.com

Recent Notes

7 min read

Golden path for my education and practice

Here is list of my projects and learning list for education, that keep me focused on my goals.

To avoid overwhelming effect, I keep this note simple as possible and hide not relevant or “future” tasks into related notes.

You can’t learn all very quickly, this is marathon, not a sprint. Strategy:

  1. Pick one topic and learn it deeply.
  2. Apply it in practice.
  3. Repeat, come back to previous projects and try to improve them!

1. Workflow and practice

My productivity workflow

  • WARN: AI in nvim workflow: Integrate AI tools (Copilot, Codeium) into Neovim carefully to enhance productivity without losing coding skills.
    • TODO: test completion, workflow to generate project
  • TODO: Practice GNU Readline keybindings

Projects

  • PROJECT: chat_oracle

    • TODO: Init UI

  • PROJECT: dev

    • TODO: Optimize for Devcontainer

  • PROJECT: nvim

    • TODO: optimize for “trixie” nvim version (old one)

  • PROJECT: VUE/flask remote game control app

    • TODO: micropython keyboard control, init hello world
    • TODO: simplest possible UI in Vue
    • TODO: remote script execution
    • TODO: keyboard/mouse listener
    • TODO: macro recorder into source code
    • TODO: pick simplest quest and automate it
    • TODO: implement some simple algorithm for fun

Software

  • TODO: MCP in Docker: Explore Model Context Protocol servers for context-aware AI interactions.
    • TODO: install hello-world MCP
    • TODO: registry of new MCP servers
  • TODO: Review my awesome software list

DSA practice

2. Science: Data Structures and Algorithms, Math

Computer Science

Programming Foundations

  • WARN: Книга: Столяров “Азы программирования” — структурное мышление.
  • TODO: Книга: Abelson and Sussman-SICP
  • TODO: Lua: Scripting for Neovim configuration and game logic.
  • TODO: Bash/Zsh: Shell scripting, automation, and dotfiles management.
  • TODO: Beej’s Guide to C Programming
  • TODO: Книга: Язык программирования C (Керниган, Ричи) — для понимания работы памяти и указателей.
  • TODO: Learning path
  • TODO: Go by Example

Mathematics

3. AI & ML Engineering

Fundamentals & ML Basics

  • WARN: Книга: Deep Learning with Python (Francois Chollet) — база по нейросетям (Keras/TensorFlow).
  • TODO: Math for AI: Векторы, матрицы, косинусное сходство (Cosine Similarity).

LLM Engineering & RAG

  • WARN: boot.dev RAG course
  • TODO: LLM APIs: OpenAI / Anthropic / Mistral / Gemini CLI.
  • TODO: RAG (Retrieval-Augmented Generation):
    • TODO: Архитектура RAG (Chunking, Embedding, Retrieval, Generation).
    • TODO: Vector Databases: pgvector (внутри Postgres), Qdrant или ChromaDB.
    • TODO: Vector Similarity Search Guide.
  • TODO: Frameworks:
    • TODO: LangChain или LlamaIndex — оркестрация цепочек.
    • TODO: DSPy — программирование промптов (тренд 2025).

Fine-Tuning & Local Models

  • TODO: Fine-tuning: Unsloth Guide (LoRA/QLoRA).
  • TODO: Inference: Запуск локальных моделей (Ollama, vLLM).
  • TODO: Prompt Engineering: Context caching, Chain-of-Thought.

4. Backend

Python

Event-Driven Architecture (EDA)

Testing & Observability

  • TODO: Testing: pytest (fixtures, markers), testcontainers (DB in docker for tests).
  • TODO: vitest/jest

API & Frameworks

  • WARN: Flask: Обзорно (Flask Mega Tutorial), чтобы понимать легаси-код.
  • TODO: FastAPI: Основной инструмент. Изучить Dependency Injection, Middleware, Background Tasks.
  • WARN: Pydantic V2: Валидация данных, .model_dump(), сериализация, работа с Settings (.env).
  • TODO: Web Protocols: HTTP/2, WebSockets, основы gRPC (важно для микросервисов).
  • TODO: WebSockets: Implement WS services (e.g., Starlette/FastAPI websockets, aiohttp).
  • TODO: FastAPI Documentation
  • TODO: Flask Documentation & Flask Mega-Tutorial
  • TODO: MDN Django Tutorial
  • TODO: FastAPI Tips: Use Lifespan State
  • TODO: Django REST Framework Quickstart - Rewrite SQLite article notes for website column.
  • TODO: Pydantic Settings: Configuration management using pydantic-settings.

Data Layer

  • WARN: Boot.dev SQL Course
  • WARN: Книга: The Art of PostgreSQL
  • WARN: Redis: Caching, pub/sub, message broker basics.
  • TODO: Asyncpg with FastAPI and Air
  • TODO: Guide to SQL JOINs
  • TODO: Indexes, transactions, isolation levels (ACID).
  • TODO: ORM:
    • TODO: Pony ORM — optional (interesting, but less common in enterprise).
    • TODO: Must Have: SQLAlchemy 2.0 (AsyncSession) — industry standard.
    • TODO: Migrations: Alembic.
  • TODO: NoSQL:

DevOps

  • WARN: Linux: Basic bash commands, scripting.
  • WARN: Linux Distributions: Understand differences between Debian-based, RHEL-based, and Arch-based systems.
  • TODO: Monitoring: Prometheus + Grafana (metrics), Sentry (error tracking).
  • TODO: Nginx: Reverse proxy, load balancing, caching, and SSL configuration.
  • TODO: systemd: Managing services, timers, and analyzing logs with journalctl.
  • TODO: CI/CD: GitLab CI/CD or GitHub Actions basics.
  • TODO: Docker & Compose: Writing optimized Dockerfiles (multi-stage builds).
  • TODO: Git: Visualization, rebase vs merge, git flow.
  • TODO: Kubernetes (K8s) Basics: Pods, Deployments, Services, and local clusters (Minikube/Kind).

5. Web development

6. Architecture, High Load and Messaging

Architecture Patterns

  • TODO: ByteByteGo - TODO: YouTube
    • TODO: pick 1 video
  • TODO: Design Patterns: Head First Design Patterns (Adapter, Factory, Strategy, Singleton).
  • TODO: Clean Architecture: 3-layer architecture (Controller → Service → Repository).
  • TODO: SOLID: Principles, especially Liskov Substitution (examples in Python).
  • TODO: Common App Patterns: Repository pattern, Service layer, Dependency Injection examples.

7. Lists & collection of courses

8. Hobby & Education

Graph View