Skip to content

Vivid Labs AI

Now Watching

The House That Remembers

The House That Remembers /AI Systems Architecture

AI Systems Architecture

Map the major parts of an AI system—from model providers and prompts to embeddings, memory, agents, and data stores—and see how they work together as one architecture.

Queue

8 videos in this course

The House That Remembers

Jump across the full chapter sequence without leaving the player.

01

AI Systems Architecture

Map the major parts of an AI system—from model providers and prompts to embeddings, memory, agents, and data stores—and see how they work together as one architecture.

02

Embeddings and Semantic Similarity

Convert textual meaning into numerical vectors, compare those vectors for semantic similarity, and understand why embeddings are the foundation of retrieval and memory.

03

Establishing an AI Provider Connection

Set up the Python environment and connect an application to an AI provider with a clean boundary for credentials, requests, and model responses.

04

Prompt Engineering

Craft focused inputs that communicate intent, context, constraints, and quality expectations so an AI agent can produce more reliable results.

05

Native Prompt Reviewer

Build a native review step that evaluates a prompt before execution, surfaces weaknesses, and turns prompt quality into an explicit part of the agent workflow.

06

The Fan-Out and Fan-In Agentic Design Pattern

Fan work out across parallel agent tasks, then fan the results back into a single synthesis step for faster and more resilient multi-stage reasoning.

07

Data Shapes

Separate semantic, episodic, and procedural data so an agent can retrieve facts, retain experiences, and follow reusable operating knowledge.

08

Tiger Cloud Architecture

Assemble a managed Postgres architecture for agent memory, operational data, and time-series signals using Tiger Cloud's AI-ready data services.