GDG San Jose DevFest event is focus on learning, building, and connecting with fellow developers and tech practitioners. They offer a variety of activities, including technical talks, workshops, and networking opportunities. This year theme is  "Responsible AI". 

We invite all to come and explore the possibilities and discuss how we can build and use AI in a fair, ethical, and responsible way.\

Common Agenda:

9:30 AM to 10:30 AM Registration

12:00 AM to 1:00 PM Lunch is provided by Google

2:30 PM to 2:45 PM Break

4:30 PM to 5:00 PM : Closing + networking + Raffle

Workshop Agenda:   Track 1

• Workshop 1: 

          10:30 AM to 12:00 PM  - Responsible AI: A Hands-On Approach by Allen Firstenberg

Description

There has been a lot of talk about the need to build AI systems that include fairness, interpretability, privacy, and safety in their core. But what models are actually doing this? And are there any tools available for developers who are using the models to help ensure responses are safe and reflect their brand name well?

We explore some tools and approaches that developers of all skill levels can use to help make sure they are using AI tools appropriately and safely.

Additional notes

The session specifically focuses on the Responsible AI components that are built into Gemini and tools that we have to tune them and how we need to handle them in our own code.

• Workshop 2: 

          1:00 PM to 2:30 PM  -  Build With AI, Fine Tuning Gemini in AI Studio and Vertex AI, a Demo by Roya Kandalan. 

Description

In this session, we dive deeper into the capabilities of Gemini beyond mere API calls. Our focus is on customizing and fine-tuning Gemini to cater to specific use cases. Finetuning enables Gemini to transform into a domain expert. We'll explore the methods and best practices for fine-tuning Gemini on AI-Studio and Vertex AI. By the end of this session, you'll have a comprehensive understanding of how to fully leverage Gemini's potential, creating a highly specialized and efficient tool tailored to your specific needs.

Additional notes

1. Attendees are required to have their laptop
2. Organizers will give GCP credit 

• Workshop 3:

        2:45 PM to 4:30 PM -  The New Paradigm in Building AI Systems by Taha Bouhsine

This workshop explores the transformative journey from traditional AI system development to the innovative paradigm powered by self-supervised learning (SSL). Attendees will learn about the limitations of the "Stone Age Paradigm," such as data unavailability, confidentiality issues, and high costs. The session will then delve into the costs and rewards of the new SSL-driven approach, highlighting how it reduces data requirements, minimizes expertise thresholds, and enhances reliability and data confidentiality.

Outline:

1. Stone Age Paradigm for Building AI Systems
2. Self-Supervised Learning
3. The New Paradigm
4. Cost and Rewards
5. Implementation of SSL (MAE, CLIP) Methods using Jax/Flax/Tensorflow

Sessions/Talks Agenda:    Track 2 

• Talk 1 

        10:30 AM to 11:15 AM -  "AI, IP, and Your Code: What Developers Need to Know" by Alla Barbalat

In a world where AI-generated code is becoming more prevalent, who owns the output? Can AI companies freely use your open-source code as training data? What are the legal ramifications when an AI system infringes upon existing intellectual property rights?

Alla will address these critical questions focused on code and IP law. The talk aims to demystify the complex legal landscape developers may find themselves navigating when using AI-generated code.

• Talk 2 

        11:15 AM to 12:00 PM  - "The Ethical Imperative of the Artificial Intelligence Explosion" by Vishnu Pendyala

Generative AI can generate demonstrably false or biased outputs, leading to real-world problems such as discrimination on the basis of protected attributes. The inner workings of AI models are often opaque, making it difficult to understand how they arrive at their outputs and hindering efforts to mitigate discrimination. Training large language models (LLMs) that constitute Generative AI can consume vast amounts of energy with excessive carbon footprint, raising sustainability concerns. Also, the ability of LLMs to generate realistic content could be exploited for misinformation campaigns or even autonomous warfare. Providing specific examples, the talk will therefore delve into the significant limitations and ethical concerns of AI such as inaccuracy and bias, lack of explainability, environmental impact, and potential for malicious use. Some of the solutions to address these issues through regulation and technology will be explained next. Governments all over the world have been enacting regulations to address issues like algorithmic discrimination and ensure consumer rights. Some specific examples of such legislation will be explained.

• Talk 3 

         1:00 PM to 1:45 PM - "Cyborgs are the future: Creating AI products to improve humanity"  by Bryce Howitson

Robots & AI are going to take your job! At least that’s the dystopian future two-thirds of Americans expect. AI & ML is now amazingly accessible and being applied across a broad spectrum of society. Unfortunately, this technology is often met with uncertainty. Fear of the robots getting it wrong. Fear of being replaced and much more set up a battle for the next century of work. But the future doesn’t have to be this way!

As technologists, it’s our job to show society the benefits of AI and to build digital tools to improve, not replace human activity. I’ll show you how. I’ll outline current perceptions of AI. I’ll share how computers have had my back. I’ll teach you to identify areas where humans fail. Then, I’ll walk through inspiring case studies of systems created to improve people.

• Talk 4

         1:45PM to 2:30PM - “Responsible AI and factuality: how to cover elections?" by Peter Danenberg

• Talk 5

         2:45PM to 3:15PM - "From Models to Morals: Responsible AI Strategies in LLMOps" Talk by Prashanth Josyula and Karanbir Singh

As artificial intelligence continues to evolve, ensuring the ethical deployment of Large Language Models (LLMs) becomes crucial. This conference session, "From Models to Morals: Responsible AI Strategies in LLMOps," will explore the integration of Responsible AI practices within the LLMOps (Large Language Model Operations) framework. The session focuses on developing strategies to address ethical challenges, including bias, privacy concerns, transparency, and governance while operationalizing LLMs effectively.

Attendees will learn about best practices for embedding Responsible AI principles throughout the LLM lifecycle, from model development to deployment and monitoring. The session will cover methods to mitigate risks associated with LLMs, such as data security and misuse, while fostering trust and accountability. Through expert talks and engaging discussions. Participants will discover how to align their LLM operations with ethical standards.

This session is ideal for AI practitioners, data scientists, and leaders who want to understand the complexities of managing AI responsibly. Join us to explore how thoughtful integration of ethical considerations into LLMOps can ensure AI technologies are both innovative and aligned with human values, paving the way for a future where AI serves the greater good responsibly.

3:15-3:30- workshop setup for IAm Remarkable.

• Talk 6

         3:30 PM to 4:30- “IAM Remarkable" workshop by Priya Gupta