500 вопросов для собеседования аналитика в Сбер
Карьерник — квиз-тренажёр в Telegram с 1500+ вопросами для собесов аналитика. SQL, Python, A/B, метрики. Бесплатно.
О компании
СберБанк — крупнейший банк России + большая tech-экосистема: Сбер Эко, Сбер Devices, СберМегамаркет, СберMarket, и много других. Tech staff 30 000+.
Analytics-команды по всем уровням и direction.
Процесс
Зависит от команды. Обычно:
- HR screening
- Technical (SQL + Python)
- Case study
- Team interview
- Final с руководителем
2-4 недели process. Иногда дольше due к security clearance.
Стек
Массивный tech:
- SQL: Oracle, Postgres, ClickHouse, Greenplum, Teradata
- Python: pandas, sklearn
- Big data: Hadoop, Spark, PySpark
- BI: Tableau, Power BI, SberBI
- ML: xgboost, CatBoost (Sber использует Yandex CatBoost)
- ETL: Airflow, Sber Airflow derivatives
SQL round
Bar moderate-high. Examples:
Middle Q1
«Customer's last 3 months avg balance».
WITH monthly_avg AS (
SELECT
customer_id,
DATE_TRUNC('month', DATE) AS month,
AVG(balance) AS avg_balance
FROM accounts
WHERE DATE BETWEEN CURRENT_DATE - INTERVAL '3 months' AND CURRENT_DATE
GROUP BY 1, 2
)
SELECT customer_id, AVG(avg_balance) AS three_month_avg
FROM monthly_avg
GROUP BY customer_id;Middle Q2
«Credit portfolio NPL rate trends MoM».
Senior Q1
«Optimize query: running 10 minutes on orders table 100M rows».
Walk through EXPLAIN, indexes, partitioning.
Python / ML
ML heavy для многих teams:
- Credit scoring
- Fraud detection
- Customer segmentation
- Next best offer
Know:
- pandas proficiency
- sklearn / xgboost / catboost
- Model evaluation (AUC, Gini)
- Feature engineering
Case study
Banking typical
«Сегмент users declining в deposit balances. Why?»
- Interest rate changes
- Competitor offers
- Cash withdrawal spike (economic stress)
- Product issue (app glitch)
«New credit card launched. Evaluate success»:
- Applications vs approvals
- Activation rates
- Spend patterns
- Default rates (long-term)
Product
Sber's huge ecosystem. Case может involve:
- SberMarket (grocery)
- SberMegamarket (marketplace)
- Sber Eco (streaming)
- Traditional banking
Different vertical, different case.
Math / stats
- Credit scoring: Gini, KS statistic
- Hypothesis testing
- Regression (для A/B)
Behavioral
Sber culture:
- Ownership
- Execution
- Learning
- Ambition
Prepare stories. Sber принципы similar к corporate values.
Security
Sber — sensitive data. Security clearance process:
- Background check
- Financial history
- Social media review
Takes time. Not scary, но prepare.
Зарплаты
- Junior: 100-150k ₽
- Middle: 180-280k ₽
- Senior: 280-400k ₽
- Lead: 400-600k ₽
- Head: 600k-1M+ ₽
Plus premiums (often 10-15% base). Yearly review.
Related teams
Many Sber companies, varying culture:
- Sber AI: R&D, research-style, highest bar
- Sber Devices: IoT, hardware-adjacent
- Sber Health: medicine, specific domain
- SberLogistika: logistics
- SberMarket: e-commerce
Apply к right fit.
Как подготовиться
SQL
Middle-strong. 50 вопросов, task collection.
Python
pandas + sklearn.
ML basics
- Logistic, random forest, xgboost
- AUC, Gini для credit
- Feature engineering
Domain
Banking products knowledge helps.
Behavioral
STAR stories. Show ownership, initiative.
Связанные темы
FAQ
Sber vs Тинькофф?
Разные culture. Sber — bigger, more corporate. Тинькофф — more startup.
ML обязательно?
Depends на team. Many teams — yes. Some не.
Remote?
Hybrid common. Some full-remote.
Готовьтесь — откройте тренажёр с 1500+ вопросами для собесов.