Newsdesk
Engineering
Data & AI
Industries
Enterprise Systems
Go-to-Market
Longform
DesignIndiaAll stories

Engineering Reading

03 October 2026

36 new posts across 7 engineering blogs — 7 worth your time.

One week of Engineering Reading, 7 stories, as published.

Practice & Craft

Why I tried to kill token billing (and why we kept it)

This post explores the tension between using token billing as an internal infrastructure mechanism and exposing it as a customer-facing pricing model. It details the lessons learned in trying to align invoice metrics with actual product value rather than raw resource costs.

Why it matters — Customer invoices should define the actual value a product delivers rather than breaking down the raw infrastructure costs incurred to create it.

Stripe · 03 October 2026 · Read the original →

Infrastructure

How we extended Apache DataFusion to execute one query across many machines

This post explains how Datadog built Distributed DataFusion to scale interactive queries across multiple machines. It details how they extended the single-node capabilities of Apache DataFusion to execute unified queries in a distributed environment.

Why it matters — Scaling interactive query engines requires extending single-node execution frameworks to coordinate and partition queries across a distributed cluster of machines.

Datadog · 03 October 2026 · Read the original →

Performance

How we built an async-aware Python profiler

This post details how Datadog built an async-aware Python profiler that preserves context across asyncio tasks. It explains the engineering decisions made to reduce profiler overhead while tracking asynchronous execution paths.

Why it matters — Profiling asynchronous applications requires tracking context across task switches rather than just thread boundaries, which must be done with minimal CPU overhead.

Datadog · 03 October 2026 · Read the original →

Data & ML

Personalization without user identity

This post explains how Airbnb personalizes search and recommendations for anonymous or logged-out users. It details how the engineering team leverages real-time proximity signals and session context instead of relying on historical user profiles.

Why it matters — Real-time proximity and session-based signals can serve as effective proxies for personalization when historical user identity is unavailable.

Airbnb · 03 October 2026 · Read the original →

Practice & Craft

Evolving our calendar assistant Reclaim to be AI-native without starting over

This post discusses the architectural evolution of Reclaim, a calendar assistant, as it integrated AI capabilities. It details how the team transitioned the system to handle ambiguous inputs and outcomes without rewriting the core deterministic scheduling engine.

Why it matters — Integrating AI into existing products is often better achieved by wrapping and evolving deterministic systems rather than completely replacing them.

Dropbox · 03 October 2026 · Read the original →

Reliability

Running multi-day AZ evacuation drills with ARC Zonal Shift

This post details how to design and run multi-day Availability Zone evacuation drills using ARC Zonal Shift. It explains how sustained stress testing helps prove application resilience and uncovers hidden architectural gaps that short-duration disaster recovery tests miss.

Why it matters — Proving true high availability requires running sustained, multi-day evacuation drills to uncover hidden dependencies that short-duration tests miss.

AWS Architecture · 03 October 2026 · Read the original →

Developer Experience

Build adaptive AI interfaces with the AG-UI protocol, agent swarms, and Nova Act on AWS

This post outlines how to build user interfaces that dynamically adapt to the variable and non-deterministic outputs of generative AI. It introduces the AG-UI protocol and details how it coordinates with agent swarms and Nova Act to reduce interface development time.

Why it matters — Non-deterministic AI outputs require dynamic, protocol-driven UI frameworks rather than static, hard-coded layouts to handle variable data structures.

AWS Architecture · 03 October 2026 · Read the original →

We’ll build distributed query engines and agent swarms before we ever fix the billing.

7 stories, every Saturday

Published here every week. Follow by RSS to get it as it lands.

← Previous issue Next issue →