Skip to content
QuantPad
FREN
Start free
Data scienceMarkets & quantM&A & PEThe LabTracks & badgesInterviewsFAQSign in
Data science

Data analyst, data scientist, data architect. One route, badge by badge.

Three jobs close enough to be confused and different enough to fail an interview on. QuantPad trains each one with graded exercises on real tables and measures how far you are from the next level.

Open the appThe Lab
The ladder

What each level is expected to do

Each level counts lessons finished, SQL and pandas exercises solved, puzzles and cards mastered. Nothing is self-declared.

  1. Data analyst

    Turns business questions into queries, dashboards and decisions.

    Joins and window functionsMetrics and cohortsA/B testspandas reshaping

  2. Data scientist

    Builds, validates and ships predictive models.

    Statistics under the modelValidation without leakagescikit-learnMonitoring drift

  3. Data architect

    Designs how data is stored, moved and trusted across the firm.

    Data modellingReplayable pipelinesStorage choicesQuality and lineage
In the Lab

A t-test on revenue, the way a team would run it.

Revenue per user includes the users who never paid. Leaving them out answers another question. The exercise checks the statistic and the p-value, not the shape of your code.

A/B test on revenue · Welch
from scipy.stats import ttest_ind

rev = (users.set_index("id")[["variant"]]
       .join(payments.groupby("user_id")["amount"].sum())
       .fillna({"amount": 0}))  # non-payers count
a, b = (rev.loc[rev.variant == v, "amount"] for v in "AB")
t = ttest_ind(a, b, equal_var=False)
result = [t.statistic, t.pvalue]
Correct · t = -1.5947, p = 0.1116
Four datasets

The same tables everywhere

Learn a table once, then query it in SQL, reshape it in pandas, model it with scikit-learn and compute on it in Julia or C++.

A shop

customers, products, orders, order_items

Joins, baskets, revenue by month, customers who never ordered.

A market

instruments, prices, traders, trades

Returns, volatility, drawdown, VWAP, correlation, P&L.

An HR base

departments, employees

Hierarchies, salaries, the classic self-join.

A product

users, events, payments

Funnels, retention, A/B tests, who pays and why.

Courses

The data tracks

Lessons that end in something you have to make work, then a badge.

6 lessons

Data Analyst

From a business question to a trustworthy number: SQL, metrics, cohorts, A/B tests and honest charts. No prerequisite.

10 lessons

Data Science

Statistics, pandas, supervised and unsupervised learning, evaluation and production: the general data scientist’s toolkit, every idea implemented by hand.

6 lessons

Data Science for Quant Research

Cleaning, causal features, validation that survives contact with reality. The half of the job that job ads list and courses skip.

6 lessons

Quant ML in Practice

The research pipeline for machine learning on market data: measure, clean, split, label, size, and judge honestly.

7 lessons

AI & Large Language Models

What is inside the models every desk now uses: training, transformers, decoding, retrieval, serving and evaluation, each with graded puzzles.

4 lessons

NLP for Finance

Turn filings, calls and news into tested signals and clean data: sentiment, event studies, de-duplication and LLM extraction.

Free · English and French

Start with a spread. Finish with an offer.

No card, no trial. Sign in and solve your first exercise in under a minute.

Open the app