MOC

AI Software & Tools

Updated 2 May 2025
mocsoftwaretoolsdevelopmentai

AI Software & Tools

The practical side. If the AI Learning Centre explains how AI works, this section covers how to build with it — the languages, frameworks, developer tools, and infrastructure that turn ideas into running systems.

This landscape moves fast. New tools appear weekly. What’s here is opinionated: we cover the things that matter, with clear guidance on what to learn first and why.

Programming Languages

The languages that power AI development. Short answer: Python dominates, TypeScript is rising for AI applications, Rust is emerging for performance-critical infrastructure.

Frameworks

The libraries you’ll actually use:

  • pytorch/ — The foundation for research and training
  • hugging-face/ — The ecosystem: models, datasets, training tools
  • langchain/ — Chains, agents, RAG orchestration
  • llamaindex/ — Data ingestion and retrieval framework
  • vercel-ai-sdk/ — Build AI features in web apps (SvelteKit, Next.js)

Developer Tools

What you work with day to day:

  • coding-assistants/ — Claude Code, Cursor, Copilot, Windsurf
  • vector-databases/ — Pinecone, Weaviate, ChromaDB, pgvector
  • inference-servers/ — Ollama, vLLM, TGI (run models locally or at scale)
  • observability/ — LangSmith, Weights & Biases, Arize (see what your AI is doing)

Infrastructure

The layer below:

  • cloud-providers/ — AWS Bedrock, GCP Vertex AI, Azure OpenAI
  • mlops/ — Deployment, monitoring, model management
  • compute/ — GPUs vs TPUs, spot instances, cost management

Protocols & Standards

The glue:

  • mcp/ — Model Context Protocol. The universal plug for AI tools.
  • openai-api-spec/ — The de facto API standard everyone implements

Where to Start

If you’re a developer coming to AI fresh:

  1. Know Python — Non-negotiable. The entire ecosystem runs on it.
  2. Get Ollama — Run models locally in one command. Experiment freely.
  3. Use a coding assistant — Claude Code, Cursor, or Copilot. Learn by doing.
  4. Build a RAG app — ChromaDB + an embedding model + Claude API. Your first real AI application.
  5. Understand AI Agents — Where everything is heading.

Go Deeper

enes