Reading path#

The collection order is intentional. Cloud infrastructure starts with the boundaries around a production application; low-level infrastructure then opens the Linux, hardware, container, and virtual-machine layers hidden beneath those services. Distributed systems adds independent clocks, partial failure, replication, and coordination, and data systems applies those mechanisms to storage engines and databases.

System design combines the earlier mechanisms into an interview and production-design method. RStack's System Design Interview Guide is a separate practice collection for Hydrocrate interview sessions, critique, and reference designs; it can be consolidated with the main system-design material later. AI inference infrastructure specializes the design method for model serving and GPUs, while harness engineering starts at the model boundary and builds the runtime that gives an agent context, tools, durable state, verification, and operating controls. Sandbox systems joins the cloud, Linux, distributed-state, and harness layers to build isolated computers for agents and untrusted code, then compares open and managed platforms. A reader can start with one specialty; each collection index names the background needed and points to slower explanations when a lower layer matters.

flowchart TB
  accTitle: Dependency map for the nine note collections
  accDescr: Cloud and low-level infrastructure supply service and machine boundaries. Distributed systems adds failure, time, replication, and coordination, which data systems applies to databases. System design combines cloud, distributed, and data mechanisms, and the interview guide applies that method in Hydrocrate practice sessions. AI inference specializes infrastructure and system design for model serving. Harness engineering starts at the model-host boundary. Sandbox systems combines cloud, Linux, distributed-state, and harness concepts for isolated execution.

  LL["Low-level infrastructure"] --> CI["Cloud infrastructure"]
  LL --> DS["Distributed systems"]
  CI --> DS
  DS --> DB["Data systems"]
  CI --> SD["System design"]
  DS --> SD
  DB --> SD
  SD --> IG["Interview guide"]
  CI --> AI["AI inference"]
  SD --> AI
  CI --> HE["Harness engineering"]
  DS --> HE
  AI -. "optional model-serving depth" .-> HE
  CI --> SS["Sandbox systems"]
  LL --> SS
  DS --> SS
  HE --> SS

The arrows show useful background, not admission gates. Start with the collection that answers your current question, then follow a dependency arrow backward when a term or failure boundary needs a slower explanation.

Collections

  1. Cloud infrastructure AWS compute, networking, identity, storage, Kubernetes control systems, queues, deployments, reliability, and recovery.
  2. Low-level infrastructure Linux boundaries, scheduling, memory, I/O, networking, containers, debugging, virtual machines, and assigned devices.
  3. Distributed systems Failure, time, coordination, replication, partitioning, distributed storage, dataflow, and recovery from first principles.
  4. Data systems Storage engines, indexes, transactions, relational databases, document stores, key-value systems, columnar files and table formats, search engines, and change data capture.
  5. System design Requirements, capacity, request paths, data systems, coordination, asynchronous work, reliability, and design cases.
  6. AI inference infrastructure Build an inference service from the first token and tensor through GPU execution, distributed serving, measurement, and safe rollout.
  7. Harness engineering Agent contracts, context, MCP and tools, durable state, verification, safety, orchestration, evaluation, and maintenance.
  8. Sandbox systems Execution boundaries, lifecycle control, warm capacity, storage, networking, identity, operations, and platform choices for untrusted code.
  9. RStack's System Design Interview Guide An independent working collection for Hydrocrate system design interview preparation.