Deep-dive guides for the MLOps toolchain

Build LLM applications from composable models, prompts, retrievers and tools. Taught at three levels — Beginner, Mid-level and Senior — each with an in-depth guide, interview prep, and practical tips.

Build stateful, controllable agents as graphs, with memory, human-in-the-loop and durable execution. Taught at three levels — Beginner, Mid-level and Senior — each with an in-depth guide, interview prep, and practical tips.

Connect LLMs to your own data: ingestion, indexing, retrieval and agentic RAG. Taught at three levels — Beginner, Mid-level and Senior — each with an in-depth guide, interview prep, and practical tips.

Build production RAG and search pipelines from modular, testable components. Taught at three levels — Beginner, Mid-level and Senior — each with an in-depth guide, interview prep, and practical tips.

Program LLMs instead of prompting them, and let optimisers tune prompts and weights for you. Taught at three levels — Beginner, Mid-level and Senior — each with an in-depth guide, interview prep, and practical tips.

Orchestrate role-based teams of AI agents that plan, delegate and collaborate on tasks. Taught at three levels — Beginner, Mid-level and Senior — each with an in-depth guide, interview prep, and practical tips.

Build multi-agent conversations and event-driven agent systems with Microsoft AutoGen. Taught at three levels — Beginner, Mid-level and Senior — each with an in-depth guide, interview prep, and practical tips.

Integrate LLMs, plugins and agents into C#, Python and Java applications with Microsoft Semantic Kernel. Taught at three levels — Beginner, Mid-level and Senior — each with an in-depth guide, interview prep, and practical tips.

Build agents with tools, handoffs, guardrails and tracing on OpenAI’s lightweight agent SDK. Taught at three levels — Beginner, Mid-level and Senior — each with an in-depth guide, interview prep, and practical tips.

Build autonomous agents on the same harness that powers Claude Code: tools, permissions, subagents and MCP. Taught at three levels — Beginner, Mid-level and Senior — each with an in-depth guide, interview prep, and practical tips.

Build, evaluate and deploy multi-agent systems with Google’s Agent Development Kit. Taught at three levels — Beginner, Mid-level and Senior — each with an in-depth guide, interview prep, and practical tips.

Build type-safe agents and structured-output LLM apps in Python with Pydantic AI. Taught at three levels — Beginner, Mid-level and Senior — each with an in-depth guide, interview prep, and practical tips.

Build lightweight code-writing agents with Hugging Face smolagents. Taught at three levels — Beginner, Mid-level and Senior — each with an in-depth guide, interview prep, and practical tips.

Connect LLM applications to tools and data through the open Model Context Protocol: build servers and clients. Taught at three levels — Beginner, Mid-level and Senior — each with an in-depth guide, interview prep, and practical tips.

Run OpenAI models in Azure: deployments, quotas, private networking and enterprise controls. Taught at three levels — Beginner, Mid-level and Senior — each with an in-depth guide, interview prep, and practical tips.

Store and search embeddings at scale with Pinecone’s managed vector database. Taught at three levels — Beginner, Mid-level and Senior — each with an in-depth guide, interview prep, and practical tips.

Run hybrid vector and keyword search with the open-source Weaviate database. Taught at three levels — Beginner, Mid-level and Senior — each with an in-depth guide, interview prep, and practical tips.

Run fast, filterable vector search with the open-source Qdrant engine. Taught at three levels — Beginner, Mid-level and Senior — each with an in-depth guide, interview prep, and practical tips.

Run billion-scale vector search with the cloud-native Milvus database. Taught at three levels — Beginner, Mid-level and Senior — each with an in-depth guide, interview prep, and practical tips.

Add embeddings search to your app in minutes with the open-source Chroma database. Taught at three levels — Beginner, Mid-level and Senior — each with an in-depth guide, interview prep, and practical tips.

Store and query embeddings inside PostgreSQL with the pgvector extension. Taught at three levels — Beginner, Mid-level and Senior — each with an in-depth guide, interview prep, and practical tips.

Build fast similarity-search indexes in-process with Meta’s FAISS library. Taught at three levels — Beginner, Mid-level and Senior — each with an in-depth guide, interview prep, and practical tips.

Fine-tune LLMs efficiently with LoRA/QLoRA (PEFT) and align them with SFT, DPO and RLHF (TRL). Taught at three levels — Beginner, Mid-level and Senior — each with an in-depth guide, interview prep, and practical tips.

Fine-tune open LLMs faster and with less GPU memory using Unsloth. Taught at three levels — Beginner, Mid-level and Senior — each with an in-depth guide, interview prep, and practical tips.

Fine-tune LLMs from a single YAML config with Axolotl: SFT, LoRA, multi-GPU and more. Taught at three levels — Beginner, Mid-level and Senior — each with an in-depth guide, interview prep, and practical tips.

Train and serve large models across many GPUs with DeepSpeed’s ZeRO and parallelism. Taught at three levels — Beginner, Mid-level and Senior — each with an in-depth guide, interview prep, and practical tips.

Organise PyTorch training code and scale it across GPUs and nodes with Lightning. Taught at three levels — Beginner, Mid-level and Senior — each with an in-depth guide, interview prep, and practical tips.

Serve LLMs with high throughput using vLLM’s PagedAttention and OpenAI-compatible server. Taught at three levels — Beginner, Mid-level and Senior — each with an in-depth guide, interview prep, and practical tips.

Serve open LLMs in production with Hugging Face Text Generation Inference. Taught at three levels — Beginner, Mid-level and Senior — each with an in-depth guide, interview prep, and practical tips.