Derick Jose is MD, Industrial AI Products within Accenture which focuses on impacting Energy, Emission and Quality outcomes for Oil and Gas, Refining, Manufacturing, Pharma, and Process Chemical industries. Prior to this, he was the cofounder at Flutura Decision Sciences which was acquired by Accenture. The product Cerebras was ranked #1 globally by Gartner peer insights which was deployed across 22 countries. He is part of multiple Industry groups related to Deeptech startups and is a visiting faculty at Stanford D School.
Join us for an insightful session where we dive deep into the real-world implementation of Generative AI, showcasing practical experiences and lessons learned from the trenches. This session will guide you through the intricacies of foundational data architecture and generative AI architecture using a real-world business case study.
Foundational Data Perspective:
In this session, you will learn about the importance of high-quality multimodal data and how to mine a variety of data, including structured, semi-structured, and unstructured formats. We will explore the concept of building data as reusable, trustworthy products and the techniques for data augmentation to make models robust with synthetic data. Additionally, you will understand how to vectorize data to enable GenAI capability and the significance of metadata and lineage for data discovery and traceability. The session will also cover data annotation and labeling, data privacy and security, and data governance policies for compliance and control. You will gain insights into the continuous monitoring of data and feedback for data quality and observability, and the concept of a self-healing data foundation with XOps, including DataOps, SecOps, and MLOps.
Generative AI Perspective:
From a generative AI perspective, the session will delve into leveraging knowledge graphs for enhanced AI capabilities and exploring architectures where AI agents interact with each other. We will discuss the development of reasoning agents capable of complex reasoning and the creation and management of digital twins for agent interactions. The integration of traditional AI techniques with generative AI will also be covered, along with an understanding of autonomous agents like BabyAGI and AutoGPT.
Advanced Concepts:
The session promises a comprehensive exploration of advanced concepts such as coordinating various AI components through a switchboard, processing and refining data for AI use in a data refinery, and designing workflows involving multiple AI agents with agentic architecture. We will emphasize the importance of ensuring ethical and responsible AI implementations and handling interactions across different data modalities for multi-modal interactions.
Read MoreManaging and scaling ML workloads have never been a bigger challenge in the past. Data scientists are looking for collaboration, building, training, and re-iterating thousands of AI experiments. On the flip side ML engineers are looking for distributed training, artifact management, and automated deployment for high performance
Read MoreManaging and scaling ML workloads have never been a bigger challenge in the past. Data scientists are looking for collaboration, building, training, and re-iterating thousands of AI experiments. On the flip side ML engineers are looking for distributed training, artifact management, and automated deployment for high performance
Read More