Как AI гик, не мог не сделать прогрев своего еще старого резюме через LLM.
Я собрал вот такой метапромт для CV review
https://gist.github.com/nikb-de/5e6669601c7e4d742b76e56ba2c4b770
Старое CV совершенно не переписано под метрики 😅, но The Roast, Red Flags detected, The Missing Evidence, Kill-Shot Questions дают прям огромную пищу для размышлений и добавлений
Оставлю только Verdict часть для кеков 🙂
Grok
HIRE at Staff (L6) with caution; the candidate demonstrates ascending trajectory through consistent platform ownership and recent LLM/anonymization work at Meta, but unfalsifiable claims like "redesigned ETL framework" miss scale metrics and failure modes, with the lowest pillar in FinOps_Unit_Economics exposing a credibility gap in cost-aware engineering. Calibrated to 10-15 YoE band in Tier 1 Tech context, where L6 expects multi-system direction-setting, this profile signals force-multiplier potential via leadership but requires interview validation on absent SLOs and AI ops maturity to avoid overplacement.
Gemini
HIRE (Strong L5 / Potential L6)
Nik demonstrates a rare and valuable "Full-Stack Data" profile: deep legacy warehousing roots (Data Vault, Oracle) successfully modernized into the cloud-native stack (Snowflake, dbt, Iceberg), and now transitioning into AI infrastructure at Meta. While some metrics are standard "resume padding" (e.g., "40% efficiency"), the trajectory from heavy lifting in banking to agile startups and finally Big Tech is coherent. The primary risk is role ambiguity—his recent past leans heavily into Engineering Management (hiring, managing 10+ reports), but he is assessed here as an IC. If he wants to code, he is a hire; if he wants to manage, he needs a different loop.
ChatGpt
DOWN-LEVEL TO L5 (Senior Data Platform Engineer) — not L6 yet.
You’ve got credible scope signals (managed teams of 7 and 10, roadmap ownership, platform building, large-cycle-time reduction for anonymization workflows), but many bullets are unfalsifiable (“overseeing function,” “near real-time data products”) with no mechanism, scale, or operational detail. The FinOps pillar is the weakest (almost no cost/unit metrics), and the AI readiness pillar reads like “worked near LLMs” rather than “built the infra that keeps LLM systems honest, cheap, and observable.”
Claude
DOWN-LEVEL TO L5 (Senior) | CONDITIONAL PASS
Twelve years of experience with demonstrated team leadership (7-10 reports) and platform-building at banking institutions, but the CV fails the 2026 AI-infrastructure bar for L6. The Meta role—ostensibly the most relevant—has the weakest evidence: "Overseeing" is not ownership, and the incomplete bullet ("providing...") suggests rushed preparation. The strongest falsifiable claims (3hr→10min load times, 3mo→2wk deployments) are from 2019-2021 Sberbank work, now heavily discounted. No SLOs, no cost-per-unit metrics, no failure modes documented anywhere. Trajectory signal is ambiguous: title progression was ascending (BI Dev → Senior → Lead) then contracted to IC at Meta—possibly Meta's flat structure, possibly scope reduction.
С ваc 100$ за карьерный коучинг 🍷
И накиньте идей, что еще добавить в метапромпт 😃
