---
title: "Discussion on \"API gateways for LLM calls\" | Hashnode"
description: "Discussion on \"API gateways for LLM calls\". Fallbacks, cooldowns, caching, and observability."
url: "https://hashnode.com/posts/llm-api-gateway/6acda33bf958039abccbabfe"
---

[Feed](https://hashnode.com/feed) /

Discussion

# [API gateways for LLM calls](https://krishparekh.hashnode.dev/llm-api-gateway?utm_source=hashnode&utm_medium=feed)

Fallbacks, cooldowns, caching, and observability.

Jul 27 · 11 min read · krishparekh.hashnode.dev  
Read the article

-   [#llm](https://hashnode.com/tag/llm)
-   [#ai](https://hashnode.com/tag/ai)
-   [#api-gateway](https://hashnode.com/tag/api-gateway)
-   [#devops](https://hashnode.com/tag/devops)
-   [#openai](https://hashnode.com/tag/openai)

## Written by

[Krish Parekh](https://hashnode.com/@KrishParekh)

@KrishParekh · Melbourne, Australia

  
[View profile](https://hashnode.com/@KrishParekh)

AI Engineer at SMEC AI · RAG, agents and voice AI

I build RAG systems, multi-agent applications and voice agents, along with the APIs and cloud infrastructure behind them. Right now that means production voice agents and ad-creative pipelines for Australian businesses. My work spans Python, FastAPI, Amazon Bedrock and Terraform, alongside Next.js and TypeScript. I write Depth First, a newsletter on production AI engineering: guardrails, evaluation, LLM gateways and retrieval.

## Responses

Markdown supported · ⌘↵ to post

No responses yet.