Why I built Lattis
I kept getting cooked on system design at Oracle, Bloomberg, and Palantir. So I built a 3D walkthrough that shows why architectures evolve as systems scale.
I fumbled the bag at places like Oracle, Bloomberg, and Palantir Technologies because I kept getting cooked on system design.
Honestly, system design is not even that hard once the basics click. The real problem is that most learning material is terrible at showing the logical progression, the why behind each architecture as a system scales to more users.
Stitching that story together from dozens of disconnected articles was rough, to say the least.
So I built Lattis to show how architecture decisions actually get made, step by step, as load grows.
The gap I kept hitting
System design interviews are not really about drawing boxes. They are about whether you get why each box shows up in a specific order:
- When do you add a cache, and what breaks if you add it too early?
- Why does fan-out come up before anyone even mentions sharding?
- What happens when one celebrity user dwarfs everyone else on the timeline?
Most prep content hands you a finished diagram and a wall of text. You memorize the picture, but the sequence of decisions never sticks. That is exactly where I kept losing rounds.
What Lattis does
Lattis is an interactive system design curriculum with 3D walkthroughs that build as you read. Click a step, watch the cache appear, and see traffic move across the system.
The idea is simple:
- One small idea per step. Twelve to fourteen steps per lesson. Each step adds a single decision instead of a lecture before anything renders.
- A scene that moves with the words. No static boxes-and-arrows screenshots. The diagram updates right as the concept shows up in the narrative.
- Tradeoffs next to the move. The reasoning stays attached to the decision that triggered it, so the order of choices never gets lost.
You walk away with a way of thinking, not a slide to memorize.
What is on the syllabus
Ten lessons, about thirty minutes each:
| Lesson | Topic | Highlights |
|---|---|---|
| Overview | System design fundamentals | 6-step process, core components, tradeoffs |
| 01 | URL Shortener | Sharding, caching, async analytics |
| 02 | Twitter (X) | Fan-out, timeline cache, the celebrity problem |
| 03 | WebSockets, offline delivery, group chat | |
| 04 | Netflix | CDN, adaptive bitrate, recommendations |
| 05 | Uber | Geospatial indexing, matching, state machines |
| 06 | Rate Limiter | Token bucket, Redis, circuit breakers |
| 07 | TicketMaster | Seat locking, virtual queues, flash sales |
| 08 | Dropbox | Chunking, dedup, conflict resolution |
| 09 | Yelp | Geo-text search, CDC, async fanout |
Who it is for
Interview prep, day job, or plain curiosity. Same deal. If big distributed systems still feel like a black box, walking through real architectures in order (with a scene that matches the story) clicks way faster than grinding random blog posts.
Early demo, want feedback
This is still an early demo. The core lessons are live and the walkthroughs work, but I am actively tuning the pacing, visuals, and depth.
If you try it and something feels confusing, too fast, or like it is missing a tradeoff you wanted spelled out, DM me on LinkedIn or email benbarrera13@gmail.com. Honest feedback is how this gets good.