Practical Utilization of AI for IT, Programming, and Development

Course overview

• Hands-On LLM Development: Gain practical expertise in developing and deploying Large Language Model (LLM)-based applications, including working with transformers, datasets, and Python-based AI models.
• Fine-Tuning and Integration: Master techniques for fine-tuning LLMs, optimizing performance, and integrating them into applications through APIs and custom solutions.
• Deployment and Monitoring: Explore strategies for secure deployment, testing, and monitoring LLMs in production environments, ensuring robust and reliable performance.
• Ethical and Responsible AI: Address ethical considerations in AI, including bias detection and mitigation, while applying responsible AI practices.

Modules

• Week 1 | Module 1: Introduction and Fundamentals
  • Understand transformer architecture, including key components like attention mechanisms and encoder-decoder structures.
  • Explore popular transformer models such as BERT, GPT, and T5, along with real-world applications.
  • Learn the capabilities, training methodologies, and differences between LLMs and traditional ML/NLP models.
  • Hands-on assignment: Use pre-trained LLMs for tasks like text generation, summarization, and sentiment analysis.
• Week 2 | Module 2: Building and Fine-Tuning LLMs
  • Learn to source, clean, and preprocess datasets, applying tokenization, padding, and batch processing techniques.
  • Fine-tune LLMs on custom datasets and optimize performance using advanced architectures like RAG, Adapter layers, and LoRA.
  • Compare fine-tuning with Retrieval-Augmented Generation (RAG) and learn techniques to mitigate overfitting, underfitting, and hallucinations.​
  • Hands-on assignment: Fine-tune an LLM and address common optimization challenges.
• Week 3 | Module 3: Leveraging LLMs Through APIs and Integration
  • Access and utilize API-based LLM services, understanding authentication, rate limits, and usage costs.
  • Build LLM-powered applications like chatbots and summarizers, leveraging evaluation metrics like BLEU, ROUGE, and Perplexity.
  • Integrate LLMs into backend and frontend applications, ensuring secure API usage and data privacy.
  • Hands-on assignment: Create a functional application using an LLM API and deploy it in an end-to-end pipeline.
• Week 4 | Module 4: Good Practices and Ethics
  • Address ethical considerations in AI, including data privacy, fairness, and bias detection/mitigation.
  • Learn responsible AI practices to ensure ethical and reliable model deployment.
  • Capstone project: Develop, deploy, and present a unique LLM-based application, showcasing your knowledge and solutions to challenges faced.

–

This program is made possible through funding by Nova Scotia's Department of Labour, Skills and Immigration's Microcredential Initiative Fund.