AI Engineer Core Track: LLM Engineering, RAG, QLoRA, Agents

Become an LLM Engineer by building 8 real applications

This is the definitive hands-on track for aspiring AI Engineers. Over 10 intensive chapters you don't just learn concepts — you build and deploy 8 real LLM applications, from an AI brochure generator and a multi-modal support agent to an autonomous deal-finding agent team.

You'll start by shipping your first LLM product while exploring the current landscape — frontier models like GPT-4o, Claude and Gemini alongside open-source models such as Llama 3 and Mistral running locally with Ollama. Then you go deeper, fast.

What you'll build and master:

• Multi-modal apps — interactive UIs with Gradio and function calling that lets LLMs take action
• Open-source generative AI — the Hugging Face ecosystem for generation, summarization and code
• The LLM showdown — evaluation frameworks (MMLU, HumanEval, custom benchmarks) to compare models on real business tasks by performance and cost
• RAG systems — vector embeddings and databases (Chroma, Pinecone, FAISS), production chunking and retrieval, plus advanced and agentic RAG with self-correcting retrieval
• Fine-tuning with QLoRA — LoRA/QLoRA, dataset curation, and advanced alignment with DPO and RLHF to build a domain model that rivals frontier models
• Autonomous multi-agent systems — ReAct and reasoning loops, LangGraph state machines, and CrewAI agent teams
• Production & career — Docker deployment, scaling, monitoring, cost optimization, guardrails, and a full-stack capstone plus portfolio guidance

Who this is for: developers with basic Python who want a project-driven path to a real AI Engineering role. By the end you'll have a portfolio of deployed LLM applications and the fine-tuning, RAG, and agent skills the market is paying top salaries for.