ROGUELIKE MODE · RUN #042
Procedurally generated runs. No undo. Real pressure.
Each run is a new scenario, new constraints, and permanent decisions. Not a quiz — a pressure test.
ACTIVE RUN — Notification System Design
Chosen
Layered Monolith — internal modules, deferred splitting
Downstream: microservice decomposition will require event contracts later
Choose one · Decision is permanent
Unlocked after Stage 2 decision · Scenario adapts to your choice
The three roguelike mechanics — applied to LLD
PROCEDURAL
Every run generates a new scenario from a seeded system — different domain, different constraints, different scale. No two runs are the same.
PERMANENT
Each decision you make is final. No backtracking, no undo, no "actually, let me try the other option." You live with your choices and see the consequences unfold.
EVALUATED
Runs end with a full breakdown — extensibility, constraint efficiency, correctness. Your score is compared across engineers who ran the same scenario type.
How a run unfolds.
Three phases, one self-contained session. Start to finish in under 30 minutes.
Start a run
A procedurally generated scenario drops you into a design problem — domain, scale, and constraints are all freshly seeded. No two runs share the same starting point.
Make design decisions
Navigate 5–10 decision points. Each choice is permanent — no going back. The scenario evolves based on what you picked earlier, so downstream stages adapt to your decisions.
See your outcome
The run ends with a full post-run evaluation — scored across extensibility, constraint efficiency, and design correctness. Compare your result to the leaderboard.
After the run
Every decision has a score. Every score has a reason.
After the final stage, you get a full breakdown of your choices — how each decision affected the overall design quality, and where the tradeoffs landed.
- Extensibility score — how well your design handles the requirement changes introduced mid-run
- Constraint efficiency — how wisely you used the budget, team, and time constraints
- Correctness — structural validity of your choices against the problem domain
RUN COMPLETE
Notification System Design
What you get
Built for pressure, scored for depth.
No undo, no safety net
Every choice commits immediately. The moment you pick an option, the scenario evolves based on it. No back button, no "actually…" — exactly like an interview.
Live resource pressure
Runs introduce budget caps, team size constraints, and timeline pressure mid-run. Your design has to hold up not just in theory, but under the constraints given.
Branching outcomes
Your early decisions shape what scenarios appear later. Choose a tightly-coupled architecture in Stage 1 and you will feel it in Stage 4 when the requirements change.
New scenario every run
Procedural generation means you cannot memorise your way through. Each run tests adaptability — the actual skill that separates interview-ready engineers.
Post-run evaluation
A scored breakdown covers extensibility, constraint efficiency, and correctness. Leaderboards are segmented by scenario type so comparisons are meaningful.
Replayability as a feature
Each new run is a different scenario. Replay is intentional — the goal is to build adaptable design instincts across a wide variety of problems, not to master one.
Who it's for
Roguelike is not for everyone. It's for these four.
Interview candidates 1–3 months out
You know the theory but freeze under real interview pressure. Roguelike puts you in that exact state repeatedly — so when the real thing arrives, it feels familiar.
Engineers who second-guess everything
If you find yourself indefinitely deferring design decisions, Roguelike forces commitment. You will learn that a committed imperfect decision outperforms paralysis every time.
Competitive learners who need stakes
If you learn better when something is on the line, the leaderboard and the permanent-decision mechanic create stakes that other practice modes simply do not have.
Engineers building daily habits
A Roguelike run fits in a focused 20-minute session. Because every run is different, the daily practice stays fresh — no two sessions are the same problem.
Ready to see how you hold up under pressure?
Start a runProcedurally generated · No two runs alike · Scored at the end
FAQ
Questions people ask
→ Explore more