Llm Engineering

This project is an educational resource and software architecture framework focused on the technical foundations of large language model engineering. It provides a collection of guides and design patterns for building and maintaining professional, scalable systems using large language models.

The resource outlines practical implementation patterns for orchestrating workflows that combine prompt engineering, model calls, and vector databases. It focuses on transforming prompt development into a structured engineering process to ensure reliable model outputs in production environments.

The covered capabilities include workflow orchestration, production prompt engineering, and the integration of vector databases to provide external context for model responses.

Features

LLM Education - Structured curriculum covering setup, Python, APIs, and deployment for learning large language model engineering.

Agentic Reasoning Frameworks - Autonomous agents are built using a reasoning-acting loop that alternates between LLM calls and tool execution.

Autonomous AI Agents - Build autonomous agents that use large language models to plan, reason, and execute multi-step tasks independently.

Autonomous Task Agents - Building AI agents that leverage LLMs for planning, reasoning, and executing multi-step tasks.

Cloud GPU Notebooks - Notebooks are designed to run on remote GPU instances, with instructions for attaching to cloud compute resources.

Curriculum Structures - Guide learners through a multi-week sequence of notebooks covering setup, Python, and deployment basics.

Weekly Module Progressions - Guided progression through weekly modules to build skills for creating an autonomous AI agent.

Large Language Model Curricula - A multi-week structured curriculum that teaches building autonomous AI agents using large language models and cloud GPUs.

Autonomous Agent Development Courses - A hands-on course guiding learners through planning, reasoning, and executing multi-step tasks with LLMs.

Jupyter Notebook Curricula - The learning material is organised as independent Jupyter notebooks that can be executed in sequence, each covering a self-contained topic.

Progressive Weekly Modules - The course content is partitioned into weekly blocks, each with incremental complexity and a final project milestone.

Cloud GPU Notebooks - Running notebooks on cloud GPUs to accelerate machine learning tasks without requiring local hardware.

Pre-Built Agent Toolkits - A pre-built set of tool functions and memory modules is provided to reduce boilerplate when constructing agents.

Ongoing Learning Programs - A weekly sequenced program covering Python, APIs, deployment, and LLM integration for agent development.

Open-Source Learning Programs - A community-driven curriculum with setup guides and contributed materials for learning LLM engineering.

ed-donnerllm_engineering

Features

Star history