Engineering blogs¶
Primary sources beat summaries: when a system-design topic comes up, read the company's own post, then compare with the textbook answer. Where a standout post is linked, it was opened and verified (September 2026). Where only "look for" topics are listed, the blog root is verified but I did not verify a specific post URL, so search the blog by the given keywords rather than trusting a guessed link. Some sites (Medium, Netflix, DoorDash, OpenAI) block automated checkers, so root links there were not machine-verified but are their well-known official addresses.
How to use
For each post ask: what was the load, what broke, what was the trade-off, what did they NOT do? Add one row to your case-study notes under docs/tracks/system-design/case-studies/.
AI labs & AI-native companies¶
| Blog | Standout posts / what to look for |
|---|---|
| Anthropic Engineering | Building effective agents · Multi-agent research system · Writing tools for agents · Effective context engineering |
| OpenAI | Look for engineering write-ups on scaling Postgres/ChatGPT infrastructure and Agents SDK/Responses API design; use the news feed already in the daily feed. |
| Hugging Face Blog | Mixture of Experts Explained · Illustrating RLHF — clear explainers plus TRL/inference-engine posts. |
| Vercel | How we built AEO tracking for coding agents (a real agent-system build); also look for AI SDK and Fluid compute posts. |
| Notion Tech | Herding elephants: sharding Postgres at Notion — a model migration write-up (shard key, dual writes, cutover). |
Consumer-scale platforms¶
| Blog | Standout posts / what to look for |
|---|---|
| Discord Engineering | How Discord Stores Trillions of Messages — Cassandra → ScyllaDB, hot partitions, request coalescing. Also look for their Elixir/Rust scale posts. |
| Netflix TechBlog | Look for: chaos engineering, Zuul/edge gateway, Cosmos/media pipelines, and data-platform posts (search the blog by these keywords). |
| Uber Engineering | Look for: Schemaless/Docstore, Ringpop, Cadence/Temporal-lineage workflows, and Michelangelo ML platform posts. |
| Airbnb Engineering | Look for: service-oriented migration, search ranking, and Minerva metrics-platform posts. |
| Pinterest Engineering | Look for: sharded MySQL/HBase evolution, Pinot-based analytics and ML serving posts. |
| Spotify Engineering | Look for: Backstage, squad/platform model, and ML platform posts. |
| Grab Tech | Look for: super-app platform, feature-store/ML and data-lake posts; good for emerging-market scale constraints. |
| Zalando Engineering | Micro Frontends: from Fragments to Renderers · also look for their API guidelines and event-driven architecture posts. |
| Canva Engineering | Look for: design-system, GraphQL and infrastructure-cost posts. |
| DoorDash Engineering | Look for: dispatch/ML, event-processing platform (Kafka/Flink) and microservices-migration posts; strong logistics relevance. |
Infrastructure, payments & developer platforms¶
| Blog | Standout posts / what to look for |
|---|---|
| Stripe Engineering | Designing robust and predictable APIs with idempotency · Online migrations at scale. |
| Cloudflare Blog | How we built Pingora · Details of the July 2, 2019 outage · Bad software deploy outage — outstanding postmortems. |
| Figma Engineering | How Figma's multiplayer technology works — CRDT-inspired collaboration design. Also look for their Postgres scaling series. |
| Shopify Engineering | Deconstructing the Monolith — modular monolith; also look for flash-sale/BFCM scale posts. |
| Meta Engineering | More details about the October 4 outage — BGP/DNS outage postmortem; also look for RocksDB, TAO and Llama infra posts. |
| Slack Engineering | Flannel: An Application-Level Edge Cache to Make Slack Scale — lazy loading and edge caching for large workspaces. |
| Dropbox Tech | Inside the Magic Pocket — exabyte-scale storage built after leaving S3. |
| LinkedIn Engineering | Look for: Kafka origins, Venice/Espresso data infra and ranking/LLM platform posts; Jay Kreps' The Log is the classic. |
| GitHub Engineering | Look for: MySQL scaling, Git infrastructure, and availability-report postmortems. |
| Datadog Engineering | Look for: Husky/Monocle time-series and event-store internals; strong on observability at scale. |
AWS Builders' Library  |
Timeouts, retries and backoff with jitter — the highest signal-to-noise cloud engineering writing. |
| Google Research Blog | Systems and ML research explained by the authors; pair with papers. |
Maersk / logistics¶
There is no dedicated Maersk engineering blog that I could verify. What exists: the open-source org MaerskTech on GitHub, the Maersk Innovation Center, and technology articles under Logistics Insights. Internal engineering knowledge lives on internal sites, not here.
Feeds worth wiring¶
Only some of these are in data/feeds.yml (mostly AI/architecture sources). If you want a system-design firehose, the weekly agent can propose adding
Cloudflare, Stripe and Netflix feeds after their RSS URLs are verified; avoid adding more than 2 per week.