Practice architecture prompts against the rubric they are judged by.
Each prompt forces clarifying questions, requirements, scale, data model, architecture, failure modes, observability, cost, and tradeoffs. The ML track adds evaluation, lineage, drift, and training / serving skew.
- Classic prompts
- 8
- ML prompts
- 6
- Rubric sections
- 10
- Reviewed prompts
- 14
Featured prompts
The first shell focuses on high-signal interview systems: queues, limiters, RAG, and model serving. Each page is reviewed, routeable, and mapped into the live sitemap.
Design a distributed rate limiter
Enforce tenant and user quotas across API gateways without turning Redis into the only thing that matters.
Design a durable job scheduler
Run delayed and recurring work with retries, backoff, leases, and a clean story for duplicate execution.
Design a RAG system
Build retrieval, generation, citation, and evaluation loops that do not collapse into a demo prompt.
Design a model serving platform
Serve models with versioning, autoscaling, canaries, and rollback paths that operators can trust.
Attempt first
The page names the problem and rubric before any model answer. The deliverable is the learner's design, not a generated essay.
Rubric visible
Each criterion names the evidence a strong answer must contain, including failure and cost reasoning.
ML systems included
RAG, model serving, feature stores, eval platforms, vector search, and distributed training are first-class prompts.