WSaaS

Grow your business and streamline your operations by harnessing the full potential of the cloud, data and AI. Subscribe to our newsletter:

[WSaaS @WSaaSHQ]

Advanced RAG Techniques | Weaviate https://t.co/cNohrAHU4B

[WSaaS @WSaaSHQ]

Making sense of AI’s contradictions: Technology infrastructure | Slalom https://t.co/4ulAkDSlNr

[Kene Oliobi @KeneOliobi]

NIST releases Dioptra for testing AI model risk

Works on only open source LLMs like Llama, Gemma, Mistral, Falcon.

https://t.co/nBqigEW70a

[WSaaS @WSaaSHQ]

Is AWS Lambda a Halo Product? https://t.co/pXJR0K9agN

[WSaaS @WSaaSHQ]

A data leader’s technical guide to scaling gen AI | Mckinsey https://t.co/PdzObhzXMk

[WSaaS @WSaaSHQ]

One year in: Lessons learned in scaling up generative AI for financial services | McKinsey https://t.co/hY1n8wSX2e

[WSaaS @WSaaSHQ]

Scaling gen AI in banking: Choosing the best operating model | McKinsey https://t.co/i5K4xHhTPW

[WSaaS @WSaaSHQ]

RT @KeneOliobi: "Humans are underrated." - Elon Musk

Human-in-the-loop is key to an effective AI/ML strategy.

Why Human Intuition Is Esse…

[WSaaS @WSaaSHQ]

Optimize Generative AI LLMs: Top 20 Hyperparameters You Need to Know | WSaaS https://t.co/CAsGd2uhXT

[WSaaS @WSaaSHQ]

Been there, doing that: How corporate and investment banks are tackling gen AI | McKinsey https://t.co/YyBWHgq0DH

[Kene Oliobi @KeneOliobi]

Optimize Generative AI LLMs: Top 20 Hyperparameters You Need to Know

#generativeai #llms #hyperparameters #deeplearning #machinelearning

https://t.co/gOsgtntfcd

[WSaaS @WSaaSHQ]

Knowledge Graph vs. Vector RAG: Benchmarking, Optimization Levers, and a Financial Analysis Example | Neo4j https://t.co/QEujQU6pKq

[WSaaS @WSaaSHQ]

A generative AI reset: Rewiring to turn potential into value in 2024 | McKinsey https://t.co/EO0kjpPZ8d

[WSaaS @WSaaSHQ]

Implementing generative AI with speed and safety | McKinsey https://t.co/PTBqW7BIlT

[WSaaS @WSaaSHQ]

Enhancing the Accuracy of RAG Applications With Knowledge Graphs | Neo4j https://t.co/rOXB7jCGFk

[WSaaS @WSaaSHQ]

Moving past gen AI's honeymoon phase: Seven hard truths for CIOs to get from pilot to scale | Gartner https://t.co/S1wzdfI0oS

[Kene Oliobi @KeneOliobi]

Event-Driven Architecture vs Message-Driven Architecture: Four Differences

When designing asynchronous / decoupled / loosely-coupled systems, many of us conflate Event-Driven Architecture with Message-Driven Architecture.

events != messages

Here are four ways to think about the differences between an Event-Driven Architecture [EDA] and a Message-Driven Architecture [MDA]:

1️⃣ EDA is Noun-Oriented. MDA is Verb-Oriented.

• EDA: Notification [Noun] that something happened. E.g., User Visited, Product Added to Cart, Order Placed. Producer is Broadcaster: Consumer chooses what to do with notification, if anything.

• MDA: Instruction [Verb] to do something. E.g., Update Inventory, Ship Order, Email Customer. Producer is Director: Consumer must follow instruction.

2️⃣ EDA is Many Consumers. MDA is Single Consumer.

• EDA: Many consumers can act on any given event / notification.

• MDA: One and only one consumer acts on any given message.

3️⃣ EDA is Dumb-Service-to-Smart Service. MDA is Smart-Service-to-Dumb-Service.

• Dumb Service: Zero to minimal business logic. E.g., Log an event [e.g., "User Visited"] without any further processing; Publish an event [e.g., "Order Placed"] to a stream without any knowledge of who will consume it.

• Smart Service: Intelligent business logic. E.g., Receive a message "Update Inventory" and perform the necessary inventory update logic; Consume a message "Email Customer" and use customer information and order details to craft and send a personalized email.

4️⃣ EDA publishes to Stream / Topic. MDA publishes to Queue.

I'll also add that you can use an event-driven system to implement a message-driven system; but generally not the other way around.

So on AWS, for instance, you can implement an event-driven architecture using:

• Amazon Managed Streaming for Apache Kafka [MSK]

• Amazon Kinesis Data Streams

• Amazon EventBridge

• Amazon Simple Notification Service [SNS]

For a message-driven architecture, I could use any of the above services. Overkill, yes, but possible.

You can implement a message-driven architecture on AWS with Amazon Simple Queue Service [SQS].

However, implementing an event-driven architecture on SQS [or attempting to do so] would be an anti-pattern.

Hope this helps!

[WSaaS @WSaaSHQ]

A generative AI reset: Rewiring to turn potential into value in 2024 | McKinsey https://t.co/EO0kjpPZ8d

[WSaaS @WSaaSHQ]

Moving past gen AI’s honeymoon phase: Seven hard truths for CIOs to get from pilot to scale | McKinsey https://t.co/S1wzdfHszk

[WSaaS @WSaaSHQ]

Navigating the Generative AI Landscape: A Strategic Blueprint for CEOs and CIOs | AWS https://t.co/GR17iepA0h