Streaming Data for AI sans Hallucination | Aerospike Summit India 2024
Learn the immense value of streaming data for AI and mechanisms with which commissioned AI provides consistent, veritable and governed outcomes. Try Aerospike Cloud for free: https://auth.control.aerospike.cloud/?utm_medium=organic-social&utm_source=youtube
In this insightful and engaging keynote from the Aerospike India Summit, Rohit Vyas, Country Head of Pre-Sales at Confluent India, explores why real-time streaming data is the backbone of reliable, governed, and high-performing AI applications.
Join Rohit as he breaks down how modern organizations can move from slow, siloed data architectures to streaming-first data platforms that power everything from Generative AI (GenAI) to fraud detection and context-aware chatbot interactions.
🧠 Key Takeaways:
- "There is no AI without data": Why real-time, contextual data is critical for both predictive and generative AI.
- The problem of hallucination in GenAI: How stale data leads to low-trust responses, and how data streaming helps solve it.
- AI architecture deep dive: How LLM applications use streaming in four key pillars: data augmentation, inference, workflow orchestration, and post-processing.
- Reverse ETL and operational feedback loops: Why insights must not only be generated but acted on in real time.
- The role of Confluent (Kafka) in modern data pipelines: Powering UPI, rideshares, fashion e-commerce, and more.
🔧 Technologies Discussed:
- Apache Kafka / Confluent Cloud
- Retrieval-Augmented Generation (RAG)
- LLMs (ChatGPT, Gemini, etc.)
- Real-time stream processing
- OLTP + OLAP convergence
- Reverse ETL
- Data governance and schema enforcement
💡 Real-World Use Cases:
- Fraud detection in banking and fintech
- Real-time customer experience in e-commerce and mobility
- Context-aware chatbots in travel and support
- LLM prompt engineering powered by current data
- Feature engineering for ML models at scale
Whether you’re a developer, architect, data scientist, or AI strategist, this session provides a comprehensive, practical, and forward-looking understanding of how real-time data streaming is the foundation for modern AI.
🧠 About the Speaker:
Rohit Vyas leads Pre-Sales at Confluent India and brings deep industry insight from helping India’s biggest digital platforms, from UPI and e-commerce to gaming and finance, build scalable, streaming-first systems.
🔗 Learn More:
Explore Aerospike: https://www.aerospike.com
Aerospike Database
https://aerospike.com/products/database/
Aerospike Documentation Guide
https://aerospike.com/docs/
Build with Aerospike
https://aerospike.com/docs/develop/
#GenerativeAI #StreamingData #Confluent
Видео Streaming Data for AI sans Hallucination | Aerospike Summit India 2024 канала Aerospike
In this insightful and engaging keynote from the Aerospike India Summit, Rohit Vyas, Country Head of Pre-Sales at Confluent India, explores why real-time streaming data is the backbone of reliable, governed, and high-performing AI applications.
Join Rohit as he breaks down how modern organizations can move from slow, siloed data architectures to streaming-first data platforms that power everything from Generative AI (GenAI) to fraud detection and context-aware chatbot interactions.
🧠 Key Takeaways:
- "There is no AI without data": Why real-time, contextual data is critical for both predictive and generative AI.
- The problem of hallucination in GenAI: How stale data leads to low-trust responses, and how data streaming helps solve it.
- AI architecture deep dive: How LLM applications use streaming in four key pillars: data augmentation, inference, workflow orchestration, and post-processing.
- Reverse ETL and operational feedback loops: Why insights must not only be generated but acted on in real time.
- The role of Confluent (Kafka) in modern data pipelines: Powering UPI, rideshares, fashion e-commerce, and more.
🔧 Technologies Discussed:
- Apache Kafka / Confluent Cloud
- Retrieval-Augmented Generation (RAG)
- LLMs (ChatGPT, Gemini, etc.)
- Real-time stream processing
- OLTP + OLAP convergence
- Reverse ETL
- Data governance and schema enforcement
💡 Real-World Use Cases:
- Fraud detection in banking and fintech
- Real-time customer experience in e-commerce and mobility
- Context-aware chatbots in travel and support
- LLM prompt engineering powered by current data
- Feature engineering for ML models at scale
Whether you’re a developer, architect, data scientist, or AI strategist, this session provides a comprehensive, practical, and forward-looking understanding of how real-time data streaming is the foundation for modern AI.
🧠 About the Speaker:
Rohit Vyas leads Pre-Sales at Confluent India and brings deep industry insight from helping India’s biggest digital platforms, from UPI and e-commerce to gaming and finance, build scalable, streaming-first systems.
🔗 Learn More:
Explore Aerospike: https://www.aerospike.com
Aerospike Database
https://aerospike.com/products/database/
Aerospike Documentation Guide
https://aerospike.com/docs/
Build with Aerospike
https://aerospike.com/docs/develop/
#GenerativeAI #StreamingData #Confluent
Видео Streaming Data for AI sans Hallucination | Aerospike Summit India 2024 канала Aerospike
Summit India 2024 AI Generative AI Streaming Data Real-Time Data Apache Kafka Confluent Data Pipelines LLM RAG Data Governance Reverse ETL Data Streaming Prompt Engineering AI Infrastructure Machine Learning Enterprise AI Data Integration Fraud Detection Real-Time Analytics Chatbots AI Hallucination OLTP OLAP Feature Engineering Confluent India Aerospike GenAI Kafka Use Cases Data Architecture
Комментарии отсутствуют
Информация о видео
11 июня 2025 г. 2:10:45
00:24:32
Другие видео канала
