Building Real-Time Gen AI Applications with SingleStore and Confluent

Discover how SingleStore and Confluent together create a powerful foundation for real-time generative AI applications. Learn how SingleStore's high-performance data platform and Confluent integrate to process and analyze streaming data in real-time. We'll explore real-world, innovative solutions and show you how SingleStore + Confluent can unlock new gen AI opportunities with your clients.

1. @yourtwitterhandle | developer.confluent.io
What are the best practices to debug client applications
(producers/consumers in general but also Kafka Streams
applications)?
Starting soon…
STARTING SOOOOON..
Starting sooooon ..
Starting soon…
Starting soon…
..Starting soon..

2. @yourtwitterhandle | developer.confluent.io
What are the best practices to debug client applications
(producers/consumers in general but also Kafka Streams
applications)?
Starting soon…
STARTING SOOOOON..
Starting sooooon ..
Starting soon…
Starting soon…

3. Copyright 2021, Confluent, Inc. All rights reserved. This document may not be reproduced in any manner without the express written permission of Confluent, Inc.
Real-time AI
Model
Config
Params,
Features
Vector Store
Object store
AI-powered
Apps
Telemetry
MLOps Pipelines
Training
Data
Output

4. Goal

Partners Tech Talks are webinars where subject matter experts from a Partner talk about a
specific use case or project. The goal of Tech Talks is to provide best practices and
applications insights, along with inspiration, and help you stay up to date about innovations
in confluent ecosystem.

5. @yourtwitterhandle | developer.confluent.io
Starting soon…
STARTING SOOOOON..
Starting sooooon ..

6. The Confluent Q2 ‘24 Launch

Accelerate your data streaming journey with partner-driven innovation

CONNECT PROCESS
Confluent Cloud for Apache Flink® Updates
Connect with Confluent Partner Program

Leverage 45+ and continuously expanding native partner
integrations to power your real-time use cases
Seamlessly Migrate to Fully Managed Connectors


Leverage custom offsets on fully managed connectors to enable
frictionless migrations, message replays, and skipping of bad records.




7 new partner additions in Q2 2024


+
Interactive Tables for Flink SQL Workspaces: Enhance productivity and streamline
workflows by exploring and visualizing Flink query results directly within the Confluent
Cloud UI
Private Networking Support on AWS (Enterprise & Dedicated LA): Ensure secure,
flexible stream processing within a private network without exposing sensitive
information to the public internet
Build with Confluent

Leverage partner built, Confluent-certified solutions to jumpstart your use case
Bring-Your-Own-Key (BYOK)
enhancements


●HSM-protected key support across
AWS, Azure and GCP
●Simplified AWS KMS policy
●BYOK API support for Google Cloud
Platform (GCP)
STREAM
Making Enterprise clusters
even more elastic


●Scale to zero when resources are not
being used
●Double peak capacity up to 2.4GBps
of throughput
●Improved cost-efficiency for
high-partition workloads
2 New Fully Managed Connectors


Unlock change data capture use cases on Amazon DynamoDB and send
real-time data to Google Cloud Functions (Gen 2) version.



Custom Connectors on Azure


Bring your own connector plugins and let Confluent Cloud manage the
underlying Connect infrastructure.



Amazon DynamoDB CDC
Source Connector
Google Cloud Functions (Gen 2)
Sink Connector

7. PROCESS
Confluent Platform for Apache Flink® (LA)
Enhanced Security with OAuth/OIDC

Streamline access to resources and data in production workloads
with OAuth, the industry standard for providing authentication
(Now Generally Available)


+
Leverage Confluent Platform for Apache Flink® (Limited
Availability) to access long-term support for on-prem or private
cloud stream processing workloads
Connect to and ingest data from external APIs and services using
the HTTP protocol and deliver it to Kafka topics in real time

Self-managed HTTP Source Connector

STREAM
CONNECT
APACHE KAFKA 3.7
Confluent Platform 7.7
Mastering Kafka security across hybrid and on-premise environments
Migrate from ZooKeeper to KRaft using CFK

Easily migrate from zookeeper to KRaft-based deployments with
Confluent for Kubernetes (CFK)
Run FIPS enabled mode with CFK
Enforce compliance by running Confluent-operator in
FIPS-enabled mode and deploying Connect, kSQL, and Schema
Registry in FIPS-enabled mode

8. @yourtwitterhandle | developer.confluent.io
Starting soon…
STARTING SOOOOON..
Starting sooooon ..

9. Confluent Perspective : AI
9

10. The Rise of Data Streaming
Data Products Versioned in git








Schema
Registry
Confluent Cloud
Consumer
Group
LLM API Gateway
LLM Instances
LLM Service
Schema Specs
Terraform
















Web Chat Agent
Vector Search
Reasoning Agent
How Confluent Works with Gen AI - Big Picture
Enforce Business Logic and Compliance Requirements with LLM Outputs
Post- Processing
Consumer Group
RAG
Airport - Physical Location
Airline Website
Bag Tracking
Cancelation/Delays
Crew Scheduling
Online Bookings
Special Offers
Loyalty Rewards
Data Governance
Field Level Tags Data Rules
Real-Time Airport, Flight, and
Price Validation
Flink SQL
Valid Answer
User Question
Kafka Topics
Producer
Connector
Connector
Connector
Producer
Connector
Crew
Customer
Flight
Gate

11. @yourtwitterhandle | developer.confluent.io
What are the best practices to debug client applications
(producers/consumers in general but also Kafka Streams
applications)?
Starting soon…
STARTING SOOOOON..
Starting sooooon ..
Starting soon…
Starting soon…

12. Building Real-Time Gen
AI Applications with
SingleStore and
Confluent
Michael Englert - Solutions Engineer

13. 13
Who Is SingleStore?
Investors
Founded
as MemSQL
Patent awarded
Three-tiered architecture
$30M funding
MemSQL 5.0
2011 2012 2013 2014 2015 2016 2017 2018 2019 2020 2021 2022 2023 2024
Vectors + semantic
search support
Rebranded to
SingleStore
Pro Max version with
ANN and compute
Raj Verma
joins as CEO
$412M in
total funding
$100M+ ARR
350+ customers

14. 14
A data platform to
transact, analyze
and contextualize
data in real time.
Row-based In-memory Distributed SQL Horizontally
scalable
Columnar Aggregate
functions
Super-fast ingest
pipelines

Vectors Semantic search Keyword search Notebooks
For on-prem, cloud and hybrid environments
Processing SQL, JSON, time-series, geospatial, etc.
✦ USPTO Patent #11,068,454

15. 15
Typical
Enterprises
Latency, inaccuracy
and cost
End
users
Enterprise
users
Slack, Google Drive,
customer support
tickets, docs, code
Applications
OLTP
database
ETL
Analytics Caching
Search Batch
Streaming
Document
database
Data
warehouse
OLAP
database
Vector
database
ETL

16. 16
Why SingleStore
SQL + JSON
Any data type

Transactions + analytics
Petabytes of enterprise data
Semantic + full-text search
Fast hybrid search

ScaleSpeed
Simplicity

17. 17
User
New AI
application
s
What is Europa?1
Response from app
based on custom prompt
6
LLM response
5
Sends question
with context
4
Vectorizes question + filter
vectors based on context
2
Match similar vectors
and rerank results
3
Analytics
Transactions
Vectors
LLM
SingleStor
e
Using SingleStore
for Context Curation

18. Copyright 2020, Confluent, Inc. All rights reserved. This document may not be reproduced in any manner without the express written permission of Confluent, Inc.
Further Information
●SingleStore Spaces
●Blog: SingleStore Brings High Performance to Vector Search
●Voyage AI Embeddings
●Docs: Tuning Vector Indexes and Queries