Testing AI Apps Workshop

Introducing our First exclusive workshop: “Testing AI Apps Workshop” Designed for professionals at all levels, Join us as we unravel the mysteries and unlock the transformative potential of this cutting-edge technology.

Workshop Duration – 8 Hours

Total Days – 2 (4 hours each day)

Date – Nov 9, 2024, and Nov 10, 2024,

Time – 5:00 PM to 9:00 PM on both days

Location: Online

Trainer: Toni Ramchandani

Toni Ramchandani

Vice President as Tech Delivery Head
at MSCI Inc.

About Trainer: Toni Ramchandani

A visionary leader and AI Research Engineer with a distinguished career that spans across multiple facets of technology and innovation. Currently serving as the Vice President as Tech Delivery Head at MSCI Inc., Toni is renowned for his expertise in Generative AI, QA, DevSecOps, Databricks, Cloud & many more. Residing in Pune, Maharashtra, India, Toni’s passion for sports, adventure, and innovation fuels his professional journey. His commitment to excellence and continuous learning makes him a valuable contributor to this workshop, where he will inspire participants with his insights and expertise. Toni has honed his skills in AI and cloud technologies, becoming a recognized figure in the industry. He is a sought after conference speaker, corporate trainer, and author, dedicated to sharing his knowledge and driving technological advancement. As an integral part of this workshop, Toni brings a wealth of knowledge, experience, and a unique perspective that will greatly benefit all participants. His journey from a curious engineering student to a leading figure in AI and technology serves as an inspiration to those looking to make a significant impact in the tech industry.

Day 1: Foundations and Fundamental Concepts (4 Hours)
1. Introduction to AI, ML, and Deep Learning (30 minutes)
Basic concepts of AI/ML.
Introduction to deep learning, neural networks, and their applications.
Overview of transformer models and their impact on modern AI.
2. Vector Databases, Embeddings, and RAGs (45 minutes)
Introduction to vector databases and their role in AI.
Understanding embeddings and their use in language models.
Overview of Retrieval-Augmented Generation (RAG) models and their applications.
3. Generative AI and LLMs (30 minutes)
Introduction to generative AI and LLMs (e.g., GPT, BERT).
Overview of LangChain and its application in building LLM-powered tools.
4. Introduction to Giskard AI (30 minutes)
Overview of Giskard’s purpose and features.
Installation and initial setup.
5. Break (15 minutes)

6. Giskard Interface and Model Integration (45 minutes)
Navigating the Giskard interface.
Integrating models (scikit-learn, TensorFlow) with Giskard.
Implementation Example: Integrate a scikit-learn model for evaluation.
7. Evaluating Tabular Models (45 minutes)
Using Giskard to evaluate tabular models.
Implementation Example: Load a tabular dataset, train a model, and evaluate it using Giskard.
8. Q&A and Wrap-Up for Day 1 (30 minutes)
Open discussion and questions from participants.
Summary of key concepts.

Day 2: Advanced Techniques and Practical Applications (4 Hours)
1. Advanced Model Evaluation Techniques (45 minutes)
Evaluating complex models like LLMs and RAGs for bias, robustness, and coherence.
Implementation Example: Evaluate a GPT model for bias using Giskard
2. Custom Tests and Transformations (45 minutes)
Creating custom tests specific to AI/ML use cases.
Using Giskard’s slicing and transformation functions
Implementation Example: Create and run custom tests on a text classification model
3. Integration with CI/CD Pipelines (45 minutes)
Automating model testing with Giskard in CI/CD workflows
Implementation Example: Integrate Giskard into a CI/CD pipeline for continuous model
evaluation
4. Break (15 minutes)
5. Collaboration, Reporting, and Advanced Integrations (45 minutes)
Collaborating within Giskard, generating reports
Integrating Giskard with external tools like MLflow, Weights & Biases, and LangChain
Implementation Example: Use MLflow to track and evaluate model versions through Giskard.
6. Real-World Case Studies and Applications (45 minutes)
Discussing case studies of Giskard in real-world AI projects, including RAG models.
Applying Giskard to participant-provided models or datasets
Implementation Example: Hands-on evaluation of a user-provided model.
7. Q&A and Workshop Wrap-Up (45 minutes)
Final questions and discussion
Summary of key learnings and next steps.
Post-Workshop
Access to Workshop Materials
Provide recorded sessions, notebooks, and additional resources.
Continued Learning and Support
Offer follow-up sessions, office hours, and continued support resources