Skip to content

Instantly share code, notes, and snippets.

@todor
Last active January 1, 2026 21:33
Show Gist options
  • Select an option

  • Save todor/3bde374e067e3dc738bb18bc4a986de7 to your computer and use it in GitHub Desktop.

Select an option

Save todor/3bde374e067e3dc738bb18bc4a986de7 to your computer and use it in GitHub Desktop.
AI Systems & LLM Engineering – Personal Learning Snapshot (Dec 2025)

AI Systems & LLM Engineering – Personal Learning Snapshot (Dec 2025)

A curated collection of resources, courses, books, and tools covering AI systems, ML, LLMs, generative AI, and agentic AI.
Snapshot for reference; my personal notes live elsewhere.


Courses


Books


Topics & Tools

  • FastAPI, uvicorn, Pydantic
  • Reasoning Models / Multi-Modal Models
  • Unsloth
  • Prompting techniques: Chain of Thought, Tree of Thought, Few-shot, Self-consistency, Reflection, ReAct
  • Generative AI / Agentic AI
  • Databricks / Spark for ML
  • Kubernetes (kagent, kgateway)
  • Vector databases & indexes: Flat, IVFFlat, HNSW, Postgres pgvector
  • RAG and Agentic RAG
  • MCP
  • LLM Orchestration: LangChain/LangGraph, CrewAI, LlamaIndex; test Agent SDKs from hyper-scalers and AI Labs
  • AI Agent & Multi-Agent Design Patterns: ReAct, Task Decomposition, Reflection, Planner-Executor, Critic-Actor, Hierarchical, Collaborative
  • Long & short memory in agentic systems + mem0
  • A2A, ACP, Internet of Agents
  • Agent Orchestration Frameworks
  • Azure AI Studio / Google Vertex AI / AWS Bedrock
  • CICD / Unit Evaluation Tests
  • liteLLM, Orq, Martian
  • LLM Observability: Langsmith SDK / Opik SDK / OpenTelemetry
  • Evaluation Techniques
  • Security: GuardrailsAI
  • OpenAI platform docs: deploy, fine-tune, embed
  • Cursor / Codex / Claude Code / ZenCoder

Short Courses

Anthropic Free Courses

  1. Prompt Engineering
  2. Building Effective Agents
  3. Best Practices for Agentic Coding
  4. The AI Fluency Framework
  5. Build with Claude
  6. Claude for Work
  7. Anthropic API Fundamentals
  8. Real World Prompting
  9. Prompt Evaluations

Posts / Repos / Resources / Papers


Agents


Transformers


Security


Subscriptions / Newsletters / Blogs

YouTube

LinkedIn

Blogs / Writers

Newsletters

Podcasts


Projects for Resume


Interviews


Cost-effective GPUs


Hands-on Projects

Smart Document QA System (RAG vs MCP)

  1. Upload documents (PDFs/Markdown)
  2. Chunk documents (e.g., LangChain TextSplitter)
  3. Create embeddings
  4. Store in vector DB
  5. Query: retrieve relevant chunks
  6. Feed context + query to LLM (OpenAI/Mistral)
  7. Return answer
  8. REST API + UI (dockerized)
  9. Add feedback loop for tuning
  10. Explore open-source models (Mistral, LLaMA2) + fine-tuning via LoRA

Other ideas:

  • Experiment with Kubeflow + GPU
  • Fine-tune a model on HuggingFace
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment