What’s included:

Program kick off starts laying the groundwork for building successful AI applications. You'll explore core AI concepts, frameworks for problem analysis and strategies for defining project success. The focus is on understanding where AI can create business value and avoiding costly project missteps by aligning technology choices with real-world needs.

‍

Learning Outcomes:

• Identify high-impact business problems where generative AI can drive measurable value.
• Define AI project success by setting business-aligned metrics, such as user engagement or operational efficiency.
• Differentiate between supervised, unsupervised and generative models to select the right approach for your business use case.
• Evaluate infrastructure requirements to support scalable and maintainable AI solutions.
• Develop a business case and project scope for your AI application, ensuring alignment with stakeholders' needs.

What’s included:

This week participants focus on the critical role of data in AI projects. You'll learn strategies for acquiring and preparing data from various sources while addressing key challenges such as data scarcity, privacy and bias. 

‍

Through hands-on activities, you'll apply techniques for data cleaning, encoding and feature engineering, ensuring your models have a solid foundation for training and evaluation.

‍

Learning Outcomes:

• Identify and integrate relevant internal and external data sources to support AI applications.
• Apply best practices in data cleaning, normalization and encoding to improve model performance.
• Mitigate risks related to data bias, privacy, and ethics by implementing responsible data handling policies.
• Explore solutions to data scarcity, including Few-Shot Learning (FSL) and leveraging large language models (LLMs).
• Develop a data pipeline that prepares your dataset for model training and testing.

What’s included:

Dive into evaluating AI models by establishing strong baseline benchmarks and exploring performance trade-offs. Learn how to select and apply evaluation metrics like precision, recall, and F1 score. 

‍

Through practical exercises, you’ll also perform error analysis to identify areas for improvement and ensure your models are effective and efficient in real-world scenarios.

‍

Learning Outcomes:

• Build and evaluate baseline models to establish a performance benchmark for AI applications.
• Select appropriate evaluation metrics (e.g., precision, recall, F1 score) based on business use cases and risks.
• Understand the trade-offs between model performance, computational efficiency and real-world constraints.
• Apply advanced error analysis techniques to improve model accuracy and mitigate risks such as bias and misclassification.
• Explore ways to optimize model performance while managing environmental and resource costs.

What’s included:

‍This week focuses on modern neural network architectures for natural language processing (NLP), including recurrent neural networks (RNNs) and transformers. 

‍

You’ll gain hands-on experience with BERT, learning how to train and fine-tune it for your own use case. The session also covers essential optimization techniques to prevent overfitting and improve model performance through regularization and feature engineering.

‍

Learning Outcomes:

• Understand the architecture and applications of neural NLP models including RNNs, LSTMs and transformers.
• Train and fine-tune a BERT model for NLP tasks such as sentiment analysis or question-answering.
• Apply optimization strategies (e.g., learning rate adjustment, dropout, early stopping) to enhance training efficiency and prevent overfitting.
• Perform advanced feature engineering and error analysis to refine model predictions and improve task-specific performance.
• Evaluate the real-world trade-offs between model complexity, training time, and resource constraints.