Registration deadline: Friday, February 26, 2027

Dates: March 11 to June 17, 2027
Schedule: Thursday evenings, 6:00 to 9:00 pm
Delivery mode: Online
Language of instruction: English
Duration: 45 hours

*Cost: $225 – see eligibility requirements*.

Description

In this course, learners will explore topics in large language models (LLMs) and enterprise AI and how they relate to secure deployment and lifecycle management.

Specifically, this course will explore the following topics:

  • Understand the end-to-end lifecycle of Large Language Models in an enterprise setting.
  • Securely deploy private open-source LLMs like Llama 3 on internal company servers.
  • Build Retrieval-Augmented Generation (RAG) pipelines to safely query proprietary data.
  • Design robust, rate-limited API endpoints for seamless internal application integration.
  • Monitor model hallucination, latency, and drift using modern observability stacks.
  • Apply scalable prompt engineering practices to ensure consistent and reliable AI outputs.
  • Optimize computational resources and GPU utilization for cost-effective AI inference.
  • Enforce strict data privacy guardrails to prevent sensitive corporate information leakage.

Prerequisites and Other Course Information

Prerequisites

  • Some understanding of Python programming (variables, functions, loops, libraries).
  • Some familiarity with REST APIs and basic cloud concepts (AWS, GCP, or Azure) – optional.
  • Some prior exposure to machine learning concepts or a foundational AI/ML course is strongly recommended.
  • Experience working in a Linux terminal environment is an asset.
  • Some understanding of version control (Git) and basic CI/CD pipelines is a plus.
  • You must ensure that you have a stable internet connection and a device capable of supporting a Teams meeting with a working microphone and camera. It is also your responsibility to verify that your device is compatible with the applications and/or software covered in the course, if applicable.

Hardware Requirements

  • CPU: Modern quad-core processor, 2.5 GHz or faster (8-core recommended for local model inference).
  • RAM: Minimum 16 GB — 32 GB strongly recommended for running LLM pipelines and containerized services simultaneously.
  • GPU: Dedicated NVIDIA GPU with 6 GB+ VRAM (e.g., RTX 3060 or higher) is highly recommended for local model fine-tuning and inference labs. Students without a compatible GPU will rely on cloud-based alternatives (covered in the course) – optional.
  • Storage: Minimum 60 GB of free SSD space — LLM model weights and datasets can be substantial in size.
  • Network: High-speed internet connection for accessing cloud GPU instances (AWS SageMaker, Google Colab Pro, or Azure ML) and pulling model weights from Hugging Face.

Software Requirements

  • Python 3.10+ with key libraries: LangChain, LlamaIndex, Hugging Face Transformers, FastAPI, Pydantic.
  • Docker Desktop — for containerized LLMOps pipelines.Git and a GitHub account — for version-controlled experiment tracking.
  • VS Code or JetBrains PyCharm (Community Edition is free).
  • Cloud account: AWS Free Tier, Google Colab (free), or Azure for Students — at least one account required.
  • Ollama or LM Studio — for running open-source LLMs locally (optional but recommended).
  • OS: Ubuntu 22.04 LTS strongly preferred; macOS 13+ (Apple Silicon supported); Windows 11 with WSL2.

Attendance and Virtual Classroom Behavior

This course requires your individual presence and your active, consistent and sustained participation in your individual work.  Your individual responsibilities are to complete the work assigned, to be online and ready to work within the first five minutes of the class.

Active Participation

While the use of the camera remains optional, active participation implies that the learner takes part in discussions, answers and asks questions, either verbally or via the chat integrated into the meeting and participates in group and/or individual exercises.  The instructor’s feedback is taken into account to validate that a learner’s participation has met the expectations for issuing the certificate.

Electronic certificates of completion will be given to students who have attended and actively participated in 80% or more of the scheduled classes.  We will no longer issue paper certificates.

Register to this course