skip to content

Frameworks

9 cheat sheets

Frameworks#

Quick-access reference for the most commonly used AI/ML frameworks — model inference with Hugging Face Transformers, LLM orchestration with LangChain, Google’s Gemini SDK, and tracing with LangSmith.

What’s in this section#

PageWhat it covers
TransformersHugging Face Transformers — pipelines, model loading, tokenizers, fine-tuning
LangChainChains, agents, memory, retrievers, document loaders
Gemini SDKGoogle Generative AI — text, vision, streaming, embeddings
LangSmithTracing, evaluation, datasets, and prompt versioning for LLM apps

Sections#

  • Transformers — Load and run pre-trained models for NLP, vision, and audio. Covers pipelines, AutoModel, tokenisation, and fine-tuning.
  • LangChain — Build LLM pipelines with LCEL chains, prompts, agents, retrievers, memory, and streaming.
  • Gemini SDK — Call Google’s Gemini models for text, multimodal input, streaming, function calling, and embeddings.
  • LangSmith — Trace, debug, evaluate, and monitor LLM apps. Covers datasets, evaluators, prompt hub, and CI integration.

transformers — Hugging Face

Load and run pre-trained models for NLP, vision, and audio with the Hugging Face Transformers library. Covers pipelines, AutoModel, tokenisation, generation, fine-tuning, and device placement.

Apr 27, 2026
python huggingface transformers llm nlp ml pytorch

google-generativeai — Gemini SDK

Call Google's Gemini models from Python for text, multimodal, streaming, chat, function calling, and embeddings. Covers the genai SDK, safety settings, file API, and async usage.

Apr 27, 2026
python google gemini llm ai multimodal api

LangChain — LLM Application Framework

Build LLM-powered pipelines with LangChain. Covers LCEL chains, chat models, prompts, output parsers, tools, agents, retrievers, memory, and streaming.

Apr 27, 2026
python langchain llm ai rag agents openai anthropic

LangSmith — LLM Observability & Evaluation

Trace, debug, evaluate, and monitor LLM applications with LangSmith. Covers tracing setup, datasets, evaluators, prompt hub, comparing runs, and CI integration.

Apr 27, 2026
python langsmith llm observability tracing evaluation langchain ai

Haystack 2.x — Pipelines for LLM Applications

Build production-grade LLM pipelines with Haystack 2.x. Covers components, the pipeline graph, indexing and querying, retrievers, generators, RAG patterns, and evaluation.

May 25, 2026
python haystack llm rag ai pipelines deepset retrievers generators

DSPy — Programmatic Prompting and Optimisation

Build LLM programs in DSPy with declarative signatures, modules, and optimisers. Covers Predict, ChainOfThought, ReAct, BootstrapFewShot, COPRO, MIPRO, MIPROv2, and inference compilation.

May 25, 2026
python dspy llm ai prompting optimisation stanford signatures modules

Semantic Kernel — Microsoft's AI Orchestration SDK

Build LLM-powered applications with Microsoft Semantic Kernel. Covers the kernel, plugins, prompt templates, planners, function calling, Kernel Memory, Python and .NET SDKs.

May 25, 2026
python dotnet csharp semantic-kernel microsoft llm ai plugins planners memory

MCP Frameworks — Servers, Clients, and Transports

Model Context Protocol (MCP) framework overview. Covers client/server architecture, stdio vs SSE vs streamable HTTP transports, FastMCP, mcp-go, the Python and TypeScript SDKs, and comparison with custom tool servers.

May 25, 2026
mcp model-context-protocol anthropic python typescript go fastmcp tools sse stdio ai

Agent Frameworks Comparison — Decision Matrix

Side-by-side comparison of LangChain, LlamaIndex, AutoGen, CrewAI, Haystack, and Semantic Kernel for building LLM-powered applications and agent systems. Covers strengths, weaknesses, and when to pick each.

May 25, 2026
ai llm agents langchain llamaindex autogen crewai haystack semantic-kernel comparison decision-matrix