benmonopoli/open-recruit / skills/tech-hiring/tech-hiring-cpp-systems-engineering
tech-hiring-cpp-systems-engineering
UnverifiedC++ systems engineers build the performance-critical infrastructure that everything else runs on: search engines, web crawlers, columnar databases, and distributed storage. Two primary contexts: (1) search and crawler infrastructure — high-throughput HTTP fetching, URL scheduling, inverted indexing at petabyte scale; (2) database internals — columnar storage engines, query execution, Raft replication, MVCC, and SIMD-optimised analytics (ClickHouse, TiFlash).
C++ Systems Engineering — Hiring Guide
How to Use This Guide
- Sourcing: Use the title synonyms and Boolean strings in each section
- Screening: Skill tiers: T1 = must-have, T2 = differentiator, T3 = rare upside
- Interviewing: Use the assessment approach to design your interview loop
- Benchmarking: Comp ranges are US market, growth-stage (Series B–D). Adjust for location, company stage, and current market
Universal Seniority Framework
Applies across all disciplines. Use as a baseline — teams vary on exact year ranges.
| Level | Typical YoE | What they own | Mentorship | Scope |
|---|---|---|---|---|
| Junior (L3 / IC1) | 0–2 years | Defined tasks, bug fixes, small features | Needs daily guidance | Self |
| Mid (L4 / IC2) | 2–5 years | Features end-to-end, moderate ambiguity | Occasional junior support | Feature / component |
| Senior (L5 / IC3) | 5–8 years | Projects, technical decisions, cross-team coordination | Active mentorship | Team / project |
| Staff (L6 / IC4) | 8–12 years | Technical strategy, cross-team systems, org-wide standards | Multiplies teams | Org |
| Principal / Distinguished | 12+ years | Company-wide technical direction, external influence | Defines engineering culture | Company |
Promotion-readiness signal: Consistently operating at the next level before the title, not just meeting current-level expectations.
C++ Systems Engineering
What This Role Does
C++ systems engineering at this level means owning the lowest layers of the stack where performance is not optional. Two distinct but related contexts:
Search & Crawler Infrastructure — Building the systems that discover, fetch, parse, and index the web at scale: URL schedulers across billions of domains, high-throughput HTTP fetching with politeness enforcement, content parsing pipelines, inverted index construction, and the distributed coordination that holds it all together.
Database Systems (ClickHouse / TiDB / TiFlash) — Building the analytical database layer: columnar storage engines, vectorised query execution, MVCC for snapshot isolation, Raft-based replication, compression codecs (LZ4, ZSTD), and SIMD-optimised operators. ClickHouse is C++ throughout. TiFlash (TiDB's analytical engine) is C++ built on ClickHouse's columnar format with Raft Learner replication.
Both contexts demand expert C++17/20, deep Linux knowledge, performance obsession backed by profiling data, and the ability to own production systems for years.
Titles & Synonyms
| Level | Common titles |
|---|---|
| Junior | Software Engineer (Systems), C++ Engineer, Systems Developer |
| Mid | C++ Engineer, Systems Engineer, Search Engineer, Database Engineer |
| Senior | Senior C++ Engineer, Senior Systems Engineer, Search Infrastructure Engineer, Database Systems Engineer |
| Staff+ | Staff Engineer, Principal Systems Engineer, Distributed Systems Architect |
| Research track | Systems Researcher, Search Relevance Engineer, Database Research Engineer |
Skill Tiers
Tier 1 — Must-have (dealbreaker if missing):
- Expert C++17 or C++20: move semantics, RAII, smart pointers, templates, concurrency primitives (std::atomic, mutexes, lock-free structures) — not just syntax, they understand what the compiler generates
- Linux systems fundamentals: TCP/IP sockets, memory-mapped files, page cache behaviour, I/O scheduling, multi-threaded programming
- Performance engineering: profiling with perf or flamegraphs, CPU cache behaviour, memory bandwidth constraints, latency vs. throughput reasoning
- Distributed systems at scale: has shipped systems handling billions of events/day or petabytes of data
- Production debugging: diagnosed memory leaks, race conditions, and performance regressions in production C++ code
Tier 2 — Differentiator:
- Search/Crawler context: inverted index internals, URL frontier management, crawl politeness (robots.txt, crawl budget), distributed deduplication (SimHash, MinHash), BM25/TF-IDF, Elasticsearch or Lucene internals
- Database context: columnar storage formats, query execution pipelines, MVCC and snapshot isolation, Raft replication, columnar compression, B-tree and LSM-tree internals, SIMD vectorisation
- RocksDB or LevelDB internals — not just usage, understands compaction, bloom filters, write amplification
- ClickHouse internals: MergeTree storage engine, materialized views, query optimizer, data types
- Custom memory allocators, arena allocation, memory pool patterns for high-throughput systems
Tier 3 — Rare Upside:
- Prior ClickHouse or TiFlash internals contributions
- SIMD intrinsics (AVX2/AVX-512) for data-parallel query execution
- Custom storage engine from scratch
- Research publications in databases (VLDB, SIGMOD, ICDE) or information retrieval (SIGIR, ECIR)
- HFT or real-time trading systems background — same performance discipline, highly transferable
Sourcing Strings
# LinkedIn X-Ray — C++ systems / infrastructure
site:linkedin.com/in/ ("C++ engineer" OR "systems engineer" OR "infrastructure engineer") (C++ OR "C plus plus") ("distributed systems" OR "search" OR "crawler" OR "database") -recruiter -talent
# Database internals specialists
site:linkedin.com/in/ (ClickHouse OR TiDB OR "TiFlash" OR "columnar" OR "OLAP") ("software engineer" OR "database engineer" OR "systems engineer") -recruiter
# Search infrastructure
site:linkedin.com/in/ ("search engineer" OR "search infrastructure" OR "crawler engineer") (C++ OR Rust OR Go) (Elasticsearch OR Lucene OR "inverted index" OR "web crawler") -recruiter
# GitHub — C++ systems engineers
language:cpp followers:>20 pushed:>2025-01-01
# GitHub — ClickHouse / database contributors
site:github.com "joined on" (ClickHouse OR TiFlash OR "columnar storage" OR "query engine") C++ followers:>15
# GitHub — search/crawler C++
site:github.com "joined on" (crawler OR "search engine" OR "inverted index") C++ followers:>15
Seniority Signals
- Junior → Mid: Moves from implementing isolated components to owning a full subsystem; starts reasoning about failure modes and production correctness, not just functionality
- Mid → Senior: Diagnoses and fixes subtle concurrency bugs in production; owns performance regressions end-to-end; their code runs for years without intervention
- Senior → Staff: Redefines the architecture of a major subsystem; identifies cross-team performance constraints; their design decisions shape how the next 10x of scale is handled
Interview & Assessment Loop
| Stage | Format | What you're assessing |
|---|---|---|
| Phone screen | Systems design discussion: "Walk me through how you'd build a distributed crawler for 1B URLs. How do you handle politeness? What fails first?" | Distributed systems thinking, production maturity |
| C++ deep dive | Real C++ problem — write a thread-safe lock-free logger, or debug a subtle race in provided code | Concurrency correctness, memory management, modern C++ |
| Systems design | Design a columnar storage engine OR a web index serving 100k QPS — breadth, trade-off reasoning | Scale thinking, storage knowledge, failure modes |
| Domain depth | "Walk me through the most performance-critical system you've owned. Show me a profiler output from it." | Real ownership, measurement discipline, production experience |
| Code review | Review a PR touching a storage or query component | Can they read unfamiliar C++? Do they find the right problems? |
Best search design question: "Design a system to crawl 10B URLs and keep them fresh within 30 days. What's your frontier strategy? How do you handle anti-bot? What breaks at 10x?" Best candidates immediately discuss crawl budget, politeness delays, prioritisation by domain authority, and distributed coordination.
Best database design question: "Design a columnar analytics engine that needs to handle 100B row inserts/day while serving sub-second queries. How do you structure writes? How do you handle compaction?" Best candidates discuss partitioning, MergeTree-style background compaction, and the write-read trade-off explicitly.
Red Flags
- "I've used ClickHouse" without understanding what MergeTree is doing internally
- Strong algorithmic problem-solver with no experience shipping systems at scale
- Writes C++ but shaky on move semantics, RAII, or what happens with a dangling reference
- Dismisses profiling: "we'll optimise if it becomes a problem" — in C++ systems, performance IS the problem
- No experience debugging race conditions or memory issues in production
- Cannot explain crawl politeness or why a crawler without robots.txt compliance is unshippable
Compensation (2025, US, Growth-Stage)
| Level | Base salary range |
|---|---|
| Junior | $120k – $160k |
| Mid | $160k – $210k |
| Senior | $195k – $275k |
| Staff | $250k – $340k |
| Principal | $310k – $420k+ |
C++ systems engineers specialising in search infrastructure or database internals command significant premiums. Expect competition from Elastic, Clickhouse Inc., PingCAP, Yandex/VK, and the major search engines. The talent pool with both C++ depth and domain-specific knowledge (crawler architecture or OLAP database internals) is very small.
Related skills
| Skill | Stars | Updated | Works with | Status | Install |
|---|---|---|---|---|---|
| writing-skillsobra/superpowersUse when creating new skills, editing existing skills, or verifying skills work before deployment | Stars: 318k | Updated | Claude CodeCodex | Unverified | Install |
| xlsxanthropics/skillsUse this skill any time a spreadsheet file is the primary input or output. This means any task where the user wants to: open, read, edit, or fix an existing .xlsx, .xlsm, .xltx, .csv, or .tsv file (e.g., adding columns, computing formulas, formatting, charting, cleaning messy data); create a new spreadsheet from scratch or from other data sources; or convert between tabular file formats. Trigger especially when the user references a spreadsheet file by name or path — even casually (like "the xlsx in my downloads") — and wants something done to it or produced from it. Also trigger for cleaning or restructuring messy tabular data files (malformed rows, misplaced headers, junk data) into proper spreadsheets. The deliverable must be a spreadsheet file. Do NOT trigger when the primary deliverable is a Word document, HTML report, standalone Python script, database pipeline, or Google Sheets API integration, even if tabular data is involved. | Stars: 196k | Updated | Claude CodeCodex | Verified | Install |
| shipping-and-launchaddyosmani/agent-skillsPrepares production launches. Use when preparing to deploy to production, or when asking what needs to be in place before shipping. Use when you need a pre-launch checklist, when setting up monitoring, when planning a staged rollout, or when you need a rollback strategy. | Stars: 118k | Updated | Claude CodeCodex | Verified | Install |
| childhoodthedotmack/claude-memBabysit a shipped release after merge, publication, or deployment. Verify published packages, production behavior, installed plugins, and late CI or user reports until the observation window is clean; use babysit for PR readiness before release. | Stars: 117k | Updated | Claude CodeCodex | Unverified | Install |