Senior Data Analytics Engineer
Drill-focused prep for a Senior Data Analytics Engineer screen — SQL creation, SQL debugging, and Python pandas LeetCode-style problems. Calibrated to a founding-analytics hire at a GPU compute marketplace. 11 chapters, two dedicated drill sets, GPU compute domain context.
If you only have a few hours
Open 00-START-HERE. Then drill 05-sql-problems and 07-pandas-problems with drill mode on. Reread 10-day-of the morning of.
Drill: SQL Problems
Creation & debugging · multiple approaches
Drill: Pandas Problems
LeetCode-style · pandas + python
Core: SQL Debugging
Broken & slow queries · 8 patterns
Section A · Orient
00
Start Here
Master index, what the screen tests, the single reframe that matters.
01
The Role, Decoded
the JD broken down — first analytics hire, GPU marketplace, founding-data-stack expectations.
02
Positioning From Scratch
Honest framing for the screen — leveraging adjacent experience, avoiding overclaiming on GPU compute domain.
Section B · SQL
03
SQL Creation Patterns
Window functions, cohort retention, funnels with time constraints, sessionization, gaps-and-islands, pivots, hierarchical queries. The patterns you compose to write SQL from a prompt.
04
SQL Debugging
Eight categories of broken SQL — NULL traps, hidden inner-join filters, duplicates from many-to-many joins, window-frame defaults, slow queries (full scans, skewed joins, missing indexes), wrong granularity.
05
SQL Problems
Ten drillable SQL problems — half creation, half debugging — with multiple approaches per problem and discussion of tradeoffs. Drill mode hides solutions.
Section C · Python & pandas
06
Pandas Fundamentals
The pandas idioms LeetCode-style screens reach for — groupby aggregates, merge gotchas, time-series ops, vectorization, the SQL↔pandas translation table, edge cases.
07
Pandas Problems
Ten drillable problems — LeetCode-style data manipulation solved in pandas with multiple approaches (vectorized vs apply vs SQL-on-pandas via DuckDB). Drill mode hides solutions.
Section D · Domain & execution
08
GPU Compute Domain
The preferred-qualification edge — GPU marketplaces, AI inference economics, utilization vs revenue, the metric vocabulary you need to sound credible with technical leadership.
09
Practice Interview Questions
~25 drillable Q&A across background, SQL, pandas, data debugging, dashboards/BI, and domain. Hide-show answers, per-question tracker.
10
Day-Of Tactics
Live-coding moves, recovery patterns, what to ask them, closing statement. Reread morning of.
Study paths
If you have 4+ days
- Day 1: 01, 02 (orient) → 03 (SQL creation patterns)
- Day 2: 04 (SQL debugging), drill 5 problems from 05
- Day 3: 06 (pandas fundamentals), drill 5 problems from 07
- Day 4: 08 (GPU domain), drill 09, reread 10
If you have 24–48 hours
01, 03 SQL creation skim, 04 SQL debugging in full, drill 5–6 SQL problems from 05, 06 pandas fundamentals, drill 4–5 pandas problems from 07, drill 09, reread 10.
If you have < 12 hours
03 window-functions section only, 04 in full, drill 3 hardest SQL problems, 06 SQL↔pandas translation table, drill 3 pandas problems, drill 09 background section, reread 10.