12 MCP, RAG, and Agents cheat sheets for AI engineers (with visuals):

1️⃣ Function calling & MCP for LLMs Before MCPs became popular, AI workflows relied on traditional Function Calling for tool access. Now, MCP is standardizing it for Agents/LLMs. The visual covers how Function Calling & MCP work under the hood. Check the thread below 👇

2️⃣ 4 stages of training LLMs from scratch This visual covers the 4 stages of building LLMs from scratch to make them practically applicable. - Pre-training - Instruction fine-tuning - Preference fine-tuning - Reasoning fine-tuning Here's my detailed thread about it 👇

3️⃣ 3 Prompting Techniques for Reasoning in LLMs This covers three popular prompting techniques that help LLMs think more clearly before they answer. - Chain of Thought (CoT) - Self-Consistency (or Majority Voting over CoT) - Tree of Thoughts (ToT) Read my detailed thread about it below 👇

4️⃣ Train LLMs using other LLMs LLMs don't just learn from raw text; they also learn from each other. - Llama 4 Scout & Maverick were trained using Llama 4 Behemoth. - Gemma 2 and 3 were trained using Gemini. The visual explains 3 popular techniques. Read the thread below 👇

5️⃣ Supervised & Reinforcement fine-tuning in LLMs This visual covers the difference between supervised fine-tuning and reinforcement fine-tuning. RFT allows you to transform any open-source LLM into a reasoning powerhouse without any labeled data. Read the thread below 👇

6️⃣ Transformer vs. Mixture of Experts in LLMs Mixture of Experts (MoE) is a popular architecture that uses different "experts" to improve Transformer models. The visual below explains how they differ from Transformers. Here's my detailed thread about it👇

7️⃣ RAG vs Agentic RAG Naive RAG retrieves once and generates once, it cannot dynamically search for more info, and it cannot reason through complex queries. Agentic RAG solves this. Check my detailed explainer thread on this👇

8️⃣ 5 popular Agentic AI design patterns Agentic behaviors allow LLMs to refine their output by incorporating self-evaluation, planning, and collaboration! This visual depicts the 5 popular design patterns for building AI agents. Check my thread on it for more info👇

9️⃣ 5 levels of Agentic AI systems Agentic systems don't just generate text; they make decisions, call functions, and even run autonomous workflows. The visual explains 5 levels of AI agency. I have linked my detailed explainer thread👇

🔟 Traditional RAG vs HyDE Questions are not semantically similar to answers, so the system may retrieve irrelevant context. In HyDE, first generate a hypothetical answer (H) to query. Then, use (H) to retrieve relevant context (C). I wrote a detailed thread about it👇

1️⃣1️⃣ RAG vs Graph RAG Answering questions that need global context is difficult with traditional RAG since it only retrieves the top-k relevant chunks. Graph RAG makes RAG more robust with graph structures. Check my detailed thread below👇

1️⃣2️⃣ KV caching KV caching is a technique used to speed up LLM inference. I have linked my detailed thread below👇

That's a wrap! If you found it insightful, reshare it with your network. Find me → @_avichawla Every day, I share tutorials and insights on DS, ML, LLMs, and RAGs.