What a Mid-level } should know

Applies algorithms for system software: B-tree for FS, hash tables with open addressing, cache-oblivious algorithms. Evaluates hardware-specific complexity.

Applies async I/O: io_uring for Linux, completion-based I/O, coroutines (C++20/Rust async). Designs event loop architectures.

Applies quality tools: sanitizers (ASAN/MSAN/TSAN/UBSAN), static analysis (clang-tidy), fuzzing for edge cases.

Applies system data structures: arena allocators, intrusive lists, radix trees, concurrent hash maps. Optimizes memory layout for cache.

Applies system patterns: Flyweight for memory optimization, Reactor for I/O multiplexing, Chain of Responsibility for packet processing.

Manages memory: custom allocators (pool, slab, arena), memory-mapped files, huge pages. Analyzes via Valgrind and perf.

Applies concurrent programming: reader-writer locks, thread pools, atomic operations, memory ordering (acquire/release/seq_cst).

Understands TCP/IP stack internals at the OS level — socket buffer tuning, congestion control algorithms, and kernel network parameter optimization. Configures TLS/SSL for custom protocol implementations. Works with low-level network primitives including raw sockets and packet crafting. Understands DNS resolution internals and TTL propagation. Debugs network issues using tcpdump, Wireshark, and eBPF network tracing.

Designs CI/CD pipelines from scratch. Configures caching for build acceleration. Uses matrix builds for multi-version testing. Sets up deployment pipelines (staging → production). Works with secrets and environment variables.

Designs system components with proper abstractions: HAL, platform abstraction, API contracts. Applies layered architecture for kernel subsystems and drivers.

Independently conducts driver and kernel module reviews. Focuses on race conditions, error handling correctness, proper kernel API usage, and memory barriers.