Master todo — Staff engineer bar + certifications + DSA + Folio

The differentiating skill is multi-team impact (ERD/RFC reviewed and adopted across teams), not another programming language.

• MAIN — Ship 1 ERD/RFC adopted by ≥2 teams (e.g. observability standard, degraded-mode playbook, agent-tool authz). Save reviewers, objections, adoption.
• End-to-end story #1: scope → design → launch → metric (full arc).
• End-to-end story #2: scope → design → launch → metric (different domain).
• End-to-end story #3: preferably the multi-team RFC, with before/after adoption.
• High collaboration: for each Staff story, name partner teams, who disagreed, how you aligned them.
• Mentorship: document 2–3 mentees — what they own now, promotion/review quality, or incident autonomy.
• Interview contribution: join company interview loops (or shadow). Log count + candidate level.
• Internal tech talk: tracing, degraded mode, or the RFC.
• External signal: architecture writeup or talk; feature it on LinkedIn (not a junior DSA tutorial).
• Process improvements: frame as “team now does X” with adoption, not one-off scripts.
• Scale metrics sheet: QPS, tenants, p99, SLO, blast radius for top résumé bullets.
• LinkedIn: Staff/Platform headline; Featured = architecture artifact; recruiter-only open-to-work for Staff/Senior Platform.
• Weekly network: 5 outreaches to target hubs (e.g. Dubai/Singapore). Ask for calibration, not a job.
• Prop-shop lane (if relevant): Python/Java platform résumé or committed C++ systems plan — not both half-done.

Keep the cert list short and serious — skip beginner pile-ons.

• SAP C_TS4CO_2023 / SAP CO (“SAP CO2”) — book, study, and sit the exam.
• CKAD (Certified Kubernetes Application Developer) — book and pass.
• If you’re leaning into AI: prioritize ML-related certifications and ship 1–2 portfolio projects that show models in production (eval, latency, cost, failure modes) — not just course badges.

Spoken practice. Pattern log on misses. Treat LeetCode Knight rank as a worthwhile milestone (signals consistent contest/performance depth beyond grinding random easies).

• Gate A — 30 spoken mediums (arrays/hash/two pointers/sliding window/BFS-DFS/recursion). 2 in 45 min.
• Pattern log: for each miss, write the wrong decision rule (not just the answer).
• Gate B — 60 total; add graphs, heaps, binary search, intervals. Spoken.
• Gate C — 90–120; company-tagged revision (Google/Meta). Timed mocks.
• Milestone: reach LeetCode Knight rank (contest / rating track).
• Jane Street lane (optional) — puzzle practice separate from LeetCode; only if applying there.
• Staff designs ×4 scored: rate limiter, notifications, task scheduler, multi-tenant authz
• Staff designs ×8: + observability pipeline, payment ledger, market-data fanout, agent tool gateway
• Complete Folio HLD problem set as Staff drills (folio.rgunkar.com HLD track)
• 1 paid/expert Staff design mock + 1 behavioral/project-retrospective mock before Google/Meta

Source: folio.rgunkar.com. Prefer core → intermediate → advanced. Expand a track below when you work it — keep the rest collapsed.