Platform & Infrastructure
Platform and infrastructure work is the 'invisible' foundation that determines whether your product can scale. The best products often win on reliability, performance, and architecture rather than feature count. Infrastructure investments require anticipating 100X scale before you need it.
The Guide
5 key steps synthesized from 5 experts.
Prioritize invisible qualities over visible features
The success of major platforms often depends on reliability, speed, privacy, and safety rather than feature count. WhatsApp won with end-to-end encryption and performance, not a feature-rich interface. Focus on the infrastructure qualities that build trust and enable scale.
Featured guest perspectives
"It wasn't the hundreds of features, it was all in the infrastructure and the platform... performance, reliability, privacy, safety, all of those things."— Asha Sharma
Abstract common capabilities into shared infrastructure
Scaling feature velocity requires identifying repetitive components and building them into a core platform architecture. Standardize capabilities like data export, filtering, or permissions at the infrastructure level so developers only focus on unique product logic.
Featured guest perspectives
"We actually stopped for the first time and say, 'What is the column like?' And we also organized all the product architecture around it... we defined what is it, and then create an infrastructure for all these shared things, making the work of adding a new column just thinking about the specific product that you want to provide."— Daniel Lereya
Set a 'doomsday clock' for scaling projects
While avoiding premature optimization is good, you need to monitor infrastructure limits and trigger scaling projects before failure. Know your breaking points (like running out of database capacity) and plan far enough ahead to avoid catastrophic outages during growth spikes.
Featured guest perspectives
"During COVID, we just couldn't scale up our infrastructure. For the longest time... we're running off even the largest instance there is for Postgres. So there's a doomsday clock... we just need to go as fast as you can to become sharding problem."— Ivan Zhao
Dedicate top talent to 100X problems
Critical infrastructure projects require dedicated teams focused on long-term architectural shifts rather than incremental features. Isolate your most talented engineers to solve 100X scale problems that will provide competitive advantage, even if it means temporarily reducing feature output.
Featured guest perspectives
"We need few of our most talented people that are now not going to contribute features anymore. We are putting them on a separate place, and let's think and solve this problem while thinking about 100X."— Daniel Lereya
Default to server-side tracking
For analytics and data infrastructure, server-side tracking is superior to client-side SDKs. It avoids data loss from ad-blockers, ensures cross-platform consistency, and reduces developer maintenance burden. Use server-side logs with user IDs as the primary source for behavioral events.
Featured guest perspectives
"The biggest mistake is setting up analytics using client side SDKs... start tracking events from your servers instead of from your clients."— Vijay
Common Mistakes
- Competing on feature count when reliability and performance matter more
- Waiting until infrastructure is failing to start scaling projects
- Building the same capabilities repeatedly instead of abstracting them into platform
- Relying on client-side tracking that loses data to ad-blockers
Signs You're Doing It Well
- New features ship faster because common capabilities are already solved
- You know your infrastructure breaking points and have projects to address them before they hit
- Enterprise customers trust you because of reliability and data residency, not just features
- Your top engineers are excited to work on foundational problems, not just shipping features
All Guest Perspectives
Deep dive into what all 5 guests shared about platform & infrastructure.
Asha Sharma
"it wasn't the hundreds of features, it was all in the infrastructure and the platform... performance, reliability, privacy, safety, all of those things."
- Prioritize reliability, speed, and end-to-end encryption in communication platforms
- Focus on data residency and availability to build trust with enterprise customers
Daniel Lereya
"We actually stopped for the first time and say, 'What is the column like?' And we also organized all the product architecture around it... we defined what is it, and then create an infrastructure for all these shared things, making the work of adding a new column just thinking about the specific product that you want to provide."
- Identify repetitive feature components and build them into a core platform architecture
- Standardize capabilities like 'export to Excel' or 'filtering' at the infrastructure level
"We need few of our most talented people that are now not going to contribute features anymore. We are putting them on a separate place, and let's think and solve this problem while thinking about 100X."
- Isolate top talent to work on long-term architectural shifts that provide a competitive edge
- Anticipate infrastructure breaking points by planning for 100X current capacity
Eli Schwartz
"If you create a categorized sitemap where you can say, 'These are all the questions on health and from the sitemap... then a search engine can navigate through the entire site, and all of the questions and answers are discoverable.'"
- Build a categorized HTML sitemap that allows bots to navigate the entire site depth
- Implement 'related content' links on every page to create a crawlable web for search engines
Ivan Zhao
"During COVID, we just couldn't scale up our infrastructure. For the longest time, Simon's really good at don't do premature optimization, so for the longest time, we Notion runs on one instance of Postgres database... we're running off even the largest instance there is for Postgres. So there's a doomsday clock... we just need to go as fast as you can to become sharding problem."
- Monitor database limits and set a 'doomsday clock' to trigger scaling projects (like sharding) before failure.
- Be prepared to halt feature development to focus entirely on critical infrastructure scaling.
Vijay
"The biggest mistake is setting up analytics using client side SDKs... start tracking events from your servers instead of from your clients."
- Default to server-side event tracking to avoid data loss from ad-blockers
- Use server-side logs with user IDs as the primary source for behavioral events
Install This Skill
Add this skill to Claude Code, Cursor, or any AI coding assistant that supports Agent Skills.
Download the skill
Download SKILL.mdAdd to your project
Create a folder in your project root and add the skill file:
.claude/skills/platform-infrastructure/SKILL.mdStart using it
Claude will automatically detect and use the skill when relevant. You can also invoke it directly:
Help me with platform & infrastructureRelated Skills
Other Engineering skills you might find useful.
Engineering Culture
A high-performance engineering culture can be built around extreme experimentation frequency and tig...
View Skill → →Managing Tech Debt
When a technical solution lacks the operational flexibility required by the business, a full rebuild...
View Skill → →Technical Roadmaps
A written strategy is essential for organizational alignment and provides a baseline that can be cri...
View Skill → →Design Engineering
The guest discusses the creation of a specific 'Design Engineering' function at Snap and Captions th...
View Skill → →