Weekend Reading #67

Weekend Reading: A weekly roundup of interesting Software Architecture and Programming articles from tech companies. Find fresh ideas and insights every weekend.

This week: data versioning patterns, Uber’s shift from batch to streaming, LinkedIn’s scalable service discovery, and Pinterest’s real-time AI for content moderation.

Data Versioning and Schema Evolution Patterns

👉 Very useful if you maintain APIs, event schemas, or long-lived databases.

In this article, we explore practical patterns for data versioning and schema evolution in NoSQL systems.

Uber: From Batch to Streaming in the Data Lake

👉 Great read if you’re modernizing data platforms or migrating to streaming.

Uber engineers describe how they moved critical pipelines from batch to streaming. The shift reduced data latency and made analytics closer to real-time.

Scalable, multi-language service discovery at LinkedIn

👉 Helpful if your organization runs a polyglot microservices architecture.

LinkedIn explains how it built a service discovery platform that works across many languages and frameworks. The system improves reliability and developer experience at scale. 

How Pinterest Built a Real‑Time Radar for Violative Content using AI

👉 Worth reading if you work on trust, safety, or real-time ML systems.

Pinterest shares how it built a real-time AI system to detect harmful or policy-violating content. The solution combines streaming data, ML models, and human review.