Data Tinkerer

What the Data Crowd Was Reading in May 2026

Tools, techniques and deep dives worth reading that I came across in May 2026.

Jun 11 • Data Tinkerer

17

2

5

May 2026

How Grab Reclaimed Hundreds of Data Engineering Hours With Multi-Agent AI

How a specialist agent system helped Grab answer data questions, check pipeline health and handle common enhancement requests without drowning engineers…

May 28 • Data Tinkerer

12

4

How Lyft Uses AI to Scale Translation Across 150+ Products

Inside the LLM pipeline targeting a 30-minute SLA for 95% of app and web translations.

May 21 • Data Tinkerer

10

3

What the Data Crowd Was Reading in April 2026

Tools, techniques and deep dives worth reading that I came across in April 2026.

May 7 • Data Tinkerer

21

1

4

April 2026

The Bitter Lesson (of Decision Making)

Why simple rules often beat human judgment over time

Apr 30 • Data Tinkerer

17

3

2

How Airtable Saved Millions by Cutting Archive Storage Costs by 100x

Airtable moved petabytes of cold log data out of MySQL and built a cheaper archive layer on S3 and Parquet without sacrificing fast queries.

Apr 23 • Data Tinkerer

15

5

How Pinterest Used Multimodal AI to Help Millions of Shoppers

Inside the multimodal AI pipeline that converted images, metadata and search behavior into scalable shopping discovery.

Apr 16 • Data Tinkerer

9

2

What the Data Crowd Was Reading in March 2026

Tools, techniques and deep dives worth reading that I came across in March 2026.

Apr 2 • Data Tinkerer

13

2

3

March 2026

How Notion Scaled AI Q&A to Millions of Workspaces

Kafka, Spark and Ray powering low-latency, high-throughput search pipelines

Mar 26 • Data Tinkerer

11

1

What the Data Crowd Was Reading in February 2026

Tools, techniques and deep dives worth reading that I came across in February 2026.

Mar 12 • Data Tinkerer

15

8

4

February 2026

How Shopify Scales Taxonomy Evolution Across 10,000+ Categories With Multi-Agent AI

From reactive manual curation to continuous taxonomy evolution grounded in merchant reality.

Feb 26 • Data Tinkerer

14

2

4

How LinkedIn Built a Pipeline That Scales to 230M Records/sec Without Breaking SLAs

From partition strategy to adaptive throttling, the playbook behind Venice’s ingestion evolution.

Feb 19 • Data Tinkerer

10

1