AI for Software Developers

Over the course of the program, participants will work in teams to build an AI application for a fictional review platform. The project will involve solving three key tasks: classifying customer reviews as positive or negative, predicting star ratings based on review text, and generating meaningful responses to customer reviews. Each week, participants will enhance their AI models through iterative development, applying machine learning techniques, natural language processing (NLP), and advanced tools like Large Language Models (LLMs) and Retrieval-Augmented Generation (RAG).

In the first week, participants will be introduced to the course structure and begin defining the core problem their AI solution will address. This foundational week sets the stage for understanding how AI models can be used to solve real business challenges by identifying the right problem to focus on. By the end of the week, participants will have a clear scope for their project and understand how to evaluate AI models.

Key Learning Objectives:

• Evaluate various open-source LLMs and understand their strengths for different use cases.

• Describe how Generative AI models work and identify which are best suited to specific challenges.

• Define a clear AI use case and project scope that addresses a real-world business need.

Task/Deliverable for Week 1:

• Create a scope of work for the application the group will build during the course, including defining the AI use case and understanding the success metrics.

In Week 2, participants will dive into acquiring and preparing the data that will be used to train their AI models. Understanding how to source and clean data is a critical step in the AI development process, as the quality of data directly impacts model performance. Participants will explore various data sources, techniques for preprocessing, and strategies for handling incomplete or noisy datasets.

Key Learning Objectives:

• Understand how to source relevant datasets and evaluate their quality.

• Apply preprocessing techniques to clean and prepare data for AI models.

• Learn strategies for handling missing or incomplete data to ensure robust model performance.

Task/Deliverable for Week 2:

• Identify a dataset related to the project use case and complete data preparation, including cleaning, normalization, and handling of missing values.

In Week 3, participants will focus on setting up baseline models for their AI projects. Establishing a baseline is essential for evaluating model performance, as it provides a reference point to measure improvements. Participants will learn how to create baseline models, define success metrics, and evaluate their models' initial performance against these metrics.

Key Learning Objectives:

• Set up baseline models and define key evaluation metrics.

• Understand the importance of an iterative approach to model improvement.

• Evaluate initial model performance and identify areas for refinement.

Task/Deliverable for Week 3:

• Build a baseline model using the prepared dataset and establish evaluation metrics to measure its performance.

In Week 4, participants will move into model development, focusing on Natural Language Processing (NLP). This week emphasizes the development and training of advanced AI models to understand and generate human language. Participants will explore how to fine-tune models to enhance performance and apply their knowledge to solve NLP challenges, such as text classification and sentiment analysis.

Key Learning Objectives:

• Develop and train NLP models using open-source LLMs.

• Fine-tune models to improve accuracy and relevance for specific use cases.

• Apply advanced NLP techniques to solve business-relevant problems.

Task/Deliverable for Week 4:

• Build and fine-tune an NLP model for a specific use case, demonstrating its ability to handle real-world data and deliver accurate results.

In Week 5, participants will delve into the intricacies of Large Language Models (LLMs), focusing on their application in solving complex business problems. This week covers the architecture of LLMs, their capabilities, and best practices for selecting and deploying them in real-world scenarios. Participants will also explore the strengths and limitations of different LLMs to ensure the right model is chosen for their specific use cases.

Key Learning Objectives:

• Understand the architecture and capabilities of various open-source LLMs.

• Evaluate and select the appropriate LLM based on the specific needs of a project.

• Learn to deploy LLMs effectively to solve complex natural language problems.

Task/Deliverable for Week 5:

• Select an LLM for your use case and demonstrate its ability to classify and generate text-based outputs for your AI project.

In Week 6, participants will explore Retrieval-Augmented Generation (RAG), a technique that enhances the capabilities of LLMs by combining them with external data retrieval. This week focuses on how RAG models can be used to improve the accuracy and relevance of responses, particularly in scenarios requiring up-to-date or domain-specific knowledge. Participants will learn how to implement and fine-tune RAG models for their AI solutions.

Key Learning Objectives:

• Understand the architecture and benefits of RAG in improving AI model outputs.

• Learn how to implement and integrate RAG into existing AI models.

• Fine-tune RAG models for specific use cases to improve response accuracy and relevance.

Task/Deliverable for Week 6:

• Implement a RAG model for your project and demonstrate how it improves the accuracy or relevance of the AI model's responses.

In Week 7, participants will focus on the art of prompting, a key technique for optimizing AI model interactions. Effective prompts can significantly enhance the performance of AI models by guiding them to generate more accurate and relevant responses. This week, participants will explore different prompting strategies and learn how to craft prompts that maximize model performance in various use cases.

Key Learning Objectives:

• Understand how different prompting techniques impact AI model performance.

• Learn to craft effective prompts to improve response quality and relevance.

• Apply advanced prompting strategies to solve specific business challenges.

Task/Deliverable for Week 7:

• Design and test a series of prompts to improve your model's ability to generate accurate and relevant outputs for your project.

In Week 8, participants will focus on fine-tuning their AI models to maximize performance. Model tuning is a crucial step in refining an AI system to meet the specific demands of a business use case. This week, participants will explore different tuning techniques, including adjusting hyperparameters and leveraging fine-tuning methods to enhance model accuracy, efficiency, and relevance.

Key Learning Objectives:

• Learn techniques for fine-tuning AI models to improve performance.

• Understand how to adjust hyperparameters to optimize model results.

• Explore advanced tuning methods to meet specific business objectives.

Task/Deliverable for Week 8:

• Fine-tune your AI model to improve accuracy and performance, and document the changes made to achieve optimal results for your project.

In Week 9, participants will be introduced to the concepts of MLOps (Machine Learning Operations), which focus on deploying and maintaining AI models in production environments. This week emphasizes the importance of a scalable and iterative approach to model deployment, ensuring AI solutions remain robust and adaptable over time. Participants will learn how to integrate AI models into production pipelines and apply best practices for monitoring and managing deployed models.

Key Learning Objectives:

• Understand the core principles of MLOps for AI model deployment and maintenance.

• Learn to implement scalable deployment pipelines for AI solutions.

• Explore best practices for monitoring, maintaining, and updating AI models in production.

Task/Deliverable for Week 9:

• Design a deployment pipeline for your AI model and demonstrate how you would monitor and maintain the model in a production environment.

In the final week, participants will present the AI solutions they have built over the course of the program. This week focuses on preparing effective presentations that clearly demonstrate the problem definition, model development, and final outcomes. Participants will reflect on their learnings, the challenges they faced, and how they overcame them. By the end of this week, participants will be able to showcase a fully functional AI solution and articulate its business impact.

Key Learning Objectives:

• Reflect on the complete AI development process and key learnings.

• Prepare and deliver an effective presentation that highlights the AI solution’s problem, approach, and outcomes.

• Demonstrate the business value and potential impact of the AI solution.

Task/Deliverable for Week 10:

• Prepare and deliver a final presentation that showcases the developed AI solution, including a discussion on the challenges faced, lessons learned, and the business impact of the project.