WeAreDevelopers Logo WeAreDevelopers
Sign in
For You LIVE
Aug 27, 2025

Reducing LLM Calls with Vector Search Patterns - Raphael De Lio (Redis)

Are massive context windows making your AI agents slow and costly? Discover how to replace direct LLM calls with Redis vector search patterns to drastically reduce compute overhead.

Pause
Mute Enter Fullscreen
#1 about 3 min

Avoiding high costs with large language model context windows

Large context models remain expensive and slow, necessitating vector search patterns for efficiency.

#2 about 3 min

Encoding textual meaning into numerical vector representations

Embedding models convert text into discrete numerical vectors to allow engines to compute semantic similarities.

#3 about 5 min

Classifying text rapidly using reference embedding generation

Pre-generating and storing embedded text references enables fast vector similarity searches for classification tasks.

#4 about 4 min

Demonstrating semantic routing thresholds with the Redis vector library

Implementing categorical text classification using a Python vector library relies on customizable distance thresholds.

#5 about 4 min

Triggering specific application tools through explicit semantic routes

Mapping user intents directly to application tools via vector databases bypasses redundant language model inferences.

#6 about 3 min

Establishing guardrails and block lists using semantic routing

Routing predefined prohibited subjects to block lists prevents irrelevant or dangerous prompts from consuming processing resources.

#7 about 2 min

Reusing language model responses through semantic database caching

Vectorizing user prompts allows applications to surface previously computed responses for semantically identical questions.

#8 about 4 min

Demonstrating semantic latency reductions using Spring AI configurations

Integrating vector databases into application frameworks dramatically reduces repeated prompt latency by retrieving cached data alongside embeddings.

#9 about 4 min

Tuning thresholds and improving accuracy with automated optimizer tools

Combining automated threshold testing with hybrid language model fallbacks refines similarity search precision continuously.

#10 about 4 min

Improving tool calling accuracy by chunking complex user prompts

Splitting verbose queries into discrete clauses ensures accurate vector matching and reliable downstream tool execution.

#11 about 4 min

Addressing caching pitfalls with specialized tools and managed services

Training specific embedding models and mixing semantic routers with caching prevents contradictory logic and enhances relevance.

Matching moments

8:02 min

Mitigating language model costs with vector search patterns

Chris Heilmann +3 · LIVE

5:11 min

Optimizing tool calling in chatbots with semantic routing

Chris Heilmann +3 · LIVE

3:13 min

Using semantic caching and pre-generated audio

Nathaniel Okenwa Nathaniel Okenwa · World Congress 2024

6:03 min

Architecting a semantic long-term memory system for LLMs

Erik Bamberg · LIVE

50 sec

Applying software engineering principles to optimize LLMs

Mete Atamel Mete Atamel · World Congress 2024

4:59 min

Optimizing server architecture for scalable large language model operations

Calvin Seward Calvin Seward · Europe 2026 Virtual

Upcoming sessions on this topic

Open session

World Congress 2026 North America

September 25, 2026 · 14:10–14:40

Stage 4

Headroom: A Context Optimization Layer for LLM Applications

Devanshi Vyas

co-founder & CTO of Headroom

Devanshi Vyas
Open session

World Congress 2026 North America

September 24, 2026 · 15:30–16:00

Stage 2

KV Cache Is Not About Speed: It's About Surviving Inference Costs

David vonThenen

AI/ML Leader | Keynote Speaker | OSS Engineer & Developer Advocate | Agentic AI, Deep Learning, Production AI | Python, Go, C++

David vonThenen
Open session

World Congress 2026 North America

September 24, 2026 · 16:10–16:40

Stage 6

Fast, Cheap, and Accurate: Optimizing LLM Inference with vLLM and Quantization

Legare Kerrison, Cedric Clyburn

Legare Kerrison
Cedric Clyburn
Open session

World Congress 2026 North America

September 24, 2026 · 17:30–18:00

Stage 6

No Single Model to Rule Them All: Building Resilient AI Agents Across Open & Closed LLMs

Emmanuel Acheampong

Senior Manager Developer Relations at Crusoe AI

Emmanuel Acheampong
Open session

World Congress 2026 North America

September 24, 2026 · 13:30–14:00

Stage 3

Understanding LLM Architectures: Inside the Design of Modern Models

Jofia Jose Prakash

Director - AI & Governance at Humanity + AI, Inc

Jofia Jose Prakash
Open session

World Congress 2026 North America

September 24, 2026 · 14:10–14:40

Stage 1

Anatomy of an AI Request: Where Latency and Cost Are Really Born

Dan Fu

VP of Kernels at Together AI

Dan Fu

Related videos

WeAreDevelopers LIVE - Vector Similarity Search Patterns for Efficiency and more thumbnail
1:05:44
LIVE • about 1 year ago

WeAreDevelopers LIVE - Vector Similarity Search Patterns for Efficiency and more

Chris Heilmann, Daniel Cranney, Raphael De Lio, Advocate At Redis

Semantic AI: Why Embeddings Might Matter More Than LLMs thumbnail
26:05
World Congress 2025 • about 1 year ago

Semantic AI: Why Embeddings Might Matter More Than LLMs

Christian Weyer

The R in RAG: Why retrieval is often the weakest link (and how to fix it) thumbnail
28:22
World Congress 2026 Europe • 3 months ago

The R in RAG: Why retrieval is often the weakest link (and how to fix it)

Tomek Porożyński

Best practices: Building Enterprise Applications that leverage GenAI thumbnail
24:24
World Congress 2025 • about 1 year ago

Best practices: Building Enterprise Applications that leverage GenAI

Damir

From minutes to seconds: Lessons learned optimizing multi-turn agentic workflows thumbnail
42:44
Europe 2026 Virtual • 3 months ago

From minutes to seconds: Lessons learned optimizing multi-turn agentic workflows

Douglas Reiser

Build RAG from Scratch thumbnail
28:04
World Congress 2024 • about 2 years ago

Build RAG from Scratch

philnash

Related articles

From learning to earning

Jobs that call for the skills explored in this talk.