Disclaimer: company names, scenarios, and incidents

Every engineering article on padho-wiki uses characters in a story to teach concepts. A pipeline is easier to remember when Riya at a fintech is fixing it at 11pm than when an unnamed engineer at an unnamed company is fixing it at an unspecified time. The story is a teaching device. The teaching is the point.

But stories tempt you to use names that sound real. A reader sees "at Razorpay, the refunds pipeline runs every fifteen minutes" and assumes someone inside Razorpay confirmed that. Nobody did. The number, the cadence, the team structure, the incident — all invented to make the explanation concrete. The same goes for every Indian company name (Flipkart, PhonePe, Zomato, Swiggy, Paytm, Hotstar, Zerodha, Cred, Dream11, Meesho, BookMyShow, Ola, Jio, Tata, Reliance, Mahindra, Infosys, TCS, Wipro, HDFC, ICICI, BigBasket, Blinkit, Zepto, Dunzo, Rapido, Ather, Practo, CureFit, Unacademy, Vedantu, PhysicsWallah, Byju's, Inshorts, Udaan, MakeMyTrip, Freshworks, Zoho, Postman, Chargebee, Groww, Upstox, Airtel, JioCinema, JioMart) and every global company name (Uber, Netflix, Lyft, DoorDash, Instacart, Airbnb, Spotify, Discord, Reddit, Pinterest, Snapchat, TikTok, Twitter, Facebook, Instagram, WhatsApp, LinkedIn, Walmart, Target, FedEx, Tesla, SpaceX, Coinbase, Robinhood, Stripe, PayPal, Square, Shopify, Booking.com, Expedia, Yelp, Notion, Figma, Canva, Salesforce, Atlassian, HubSpot, Twilio, Zendesk, Intercom, ByteDance, Alibaba, Tencent, Baidu) you may once have seen on this site.

To remove the ambiguity, all such mentions have been replaced with hypothetical company names across every engineering article. If a story now reads "at PaisaBridge, the refunds pipeline runs every fifteen minutes" — PaisaBridge is invented. The pipeline is invented. The number is invented. The engineer is invented.

What this means concretely

• Company names are hypothetical. Any company you read about in an engineering article on this site does not exist. Names that sound like real companies — Indian or global — were chosen to feel concrete, not to refer to the real company. Even where a name resembles a real one, treat the company in the story as fictional.
• Engineers are hypothetical. Riya, Aditi, Aman, Priya, Karan — none of them work at the company in the story. They are characters built to make a concept memorable.
• Numbers are hypothetical. Throughput numbers, cost numbers, headcount, traffic patterns, peak-hour multipliers, refund volumes, request rates — all chosen to be plausible for a teaching example, not measured from any real production system.
• Incidents are hypothetical. Outages, post-mortems, on-call stories, "the time the pipeline broke during Diwali" — invented for pedagogy. Any resemblance to real incidents at real companies is coincidental.
• Scaling-up sketches are hypothetical. When an article says "the production version at <Company> would look like this," the production version sketched is not what that real company runs. It is what a textbook example of "production-shape" code might look like, attributed to a name to make it concrete.
• External links may be inaccurate. Some articles cite engineering blog posts at company URLs. These links may be broken, may point to posts that do not exist, or may point to posts that do not say what the article claims they say. Treat any external citation as a starting point for your own search, not a verified reference.

In short: any company-shaped fact you see on this site is hypothetical, even when it resembles something real. The technical concepts (Kafka semantics, Iceberg manifests, Spanner's TrueTime, S3's consistency model) are real and accurate to the best of our knowledge. The stories around them are invented.

Why we tell stories at all

Without a story, an article on append-only logs reads like a manual. With a story, it reads like a problem someone is actually solving. We chose stories. The cost is that we have to be loud about the stories being fiction, which is the point of this page.

If you want the canonical source for a technical claim — Kafka's exactly-once semantics, Iceberg's commit protocol, BigQuery's slot model, Snowflake's micro-partitions — go to the original paper, the official documentation, or the open-source repository. We try to point you there. We do not always succeed.

Technical product names

A separate, narrow exception: where a product name is also a company name (Snowflake, Datadog, Confluent, Cloudflare, GitHub, Slack, MongoDB, Oracle, Databricks), we may keep the product name when it refers to the product as a teaching subject — e.g., "Snowflake's micro-partition format" — because renaming the product would make the article useless to a reader trying to learn that product. When the same name appears as a company actor in a story, it is fictionalised like any other.

Reporting an issue

If you see a place where a real company name slipped through, or a citation that should not be trusted, please tell us — the email is on the home page. We will fix it.

— the padho-wiki team