DEV Community

Architecture

The fundamental structures of a software system.

Posts

đź‘‹ Sign in for the ability to sort posts by relevant, latest, or top.
How I Built a Cost-Aware Layer for LLM APIs: Caching, Routing, and Fallback

How I Built a Cost-Aware Layer for LLM APIs: Caching, Routing, and Fallback

Comments
7 min read
AI Agents should be able to share context, not just messages

AI Agents should be able to share context, not just messages

Comments
3 min read
MobileTopUP: Designing a Reliable Recharge Transaction Workflow

MobileTopUP: Designing a Reliable Recharge Transaction Workflow

Comments 1
5 min read
Structured Output Is an API: Evolving Gemini Schemas Without Breaking Your Agent

Structured Output Is an API: Evolving Gemini Schemas Without Breaking Your Agent

8
Comments
3 min read
Why CLAUDE.md Is Not Enough to Protect Your Architecture

Why CLAUDE.md Is Not Enough to Protect Your Architecture

Comments
4 min read
Giving AI Coding Agents Context Without Giving Them Your Entire Codebase

Giving AI Coding Agents Context Without Giving Them Your Entire Codebase

Comments 5
5 min read
How to Build an AI-Ready Customer Profile

How to Build an AI-Ready Customer Profile

1
Comments 1
8 min read
Giving AI Agents a Deterministic Clock (Zero LLM in the Loop)

Giving AI Agents a Deterministic Clock (Zero LLM in the Loop)

Comments
3 min read
Who Calls Me, Not Whom I Call

Who Calls Me, Not Whom I Call

Comments 1
10 min read
I put reminders, contacts and recipes in the same Postgres table. Here is the SQL, and what it cost.

I put reminders, contacts and recipes in the same Postgres table. Here is the SQL, and what it cost.

Comments
5 min read
Building a Modern Data Lakehouse on AWS: S3, Iceberg, Glue, Athena, and Lake Formation

Building a Modern Data Lakehouse on AWS: S3, Iceberg, Glue, Athena, and Lake Formation

Comments
6 min read
Your agent's cost problem isn't the model. It's the steps you never measured.

Tracks token spend by step, tool, and model tuple

Your agent's cost problem isn't the model. It's the steps you never measured.

10
Comments 13
4 min read
Mycelium – Sub-10ms Semantic Tool Routing for AI Agents (No LLM overhead)

Mycelium – Sub-10ms Semantic Tool Routing for AI Agents (No LLM overhead)

2
Comments 4
2 min read
Why marketplace accounting becomes so hard

Why marketplace accounting becomes so hard

1
Comments
15 min read
Custodial ETH withdrawals: the nonce is a database row, not a node call

Custodial ETH withdrawals: the nonce is a database row, not a node call

Comments
7 min read
đź‘‹ Sign in for the ability to sort posts by relevant, latest, or top.