Intuit OA 2026 | DS Take-home + SQL Case Full Walkthrough

Intuit (TurboTax / QuickBooks / Mint / Credit Karma) OA is a "take-home + SQL case + Karat live coding" trio. DS and SDE tracks differ in question style but both lean real business decision-making over pure algorithms. Below: 2026 spring-cycle latest signals + 5-day sprint.

Intuit OA Snapshot (2026)

Line 1: Take-home Case (DS Track)

Prompt

A TurboTax user-behavior CSV (100K rows). Tasks:

1. EDA + data-quality assessment
2. Hypothesis: "why is H1 filing-conversion 8% higher than H2?"
3. Design an experiment to validate

Response Framework

import pandas as pd

def eda(df):
    print(df.shape, df.dtypes)
    print(df.isnull().sum().sort_values(ascending=False).head(10))
    print(df.describe(include='all').T)
    for col in ['half_year', 'is_returning', 'device', 'tax_year']:
        rates = df.groupby(col)['filed'].mean()
        print(col, rates.to_dict())
    return df

Intuit grades business-conclusion soundness + experiment executability, not just code. Surface H1/H2 user-mix differences early in EDA.

Line 2: SQL Business Case

WITH active_months AS (
    SELECT
        s.user_id,
        s.plan,
        DATE_PART('month', AGE(COALESCE(s.churned_at, CURRENT_DATE), s.started_at))
            + DATE_PART('year', AGE(COALESCE(s.churned_at, CURRENT_DATE), s.started_at)) * 12
            AS months_active
    FROM subscription s
    WHERE s.started_at >= DATE '2025-01-01'
), revenue AS (
    SELECT user_id, SUM(amount) AS total_rev
    FROM charge
    WHERE ts >= DATE '2025-01-01'
    GROUP BY user_id
), per_user_ltv AS (
    SELECT
        a.plan,
        a.user_id,
        CASE
            WHEN a.months_active = 0 THEN COALESCE(r.total_rev, 0)
            ELSE COALESCE(r.total_rev, 0) / a.months_active * LEAST(a.months_active, 12)
        END AS ltv12
    FROM active_months a
    LEFT JOIN revenue r ON a.user_id = r.user_id
)
SELECT plan, AVG(ltv12) AS ltv_avg, COUNT(*) AS n
FROM per_user_ltv
GROUP BY plan
ORDER BY ltv_avg DESC;

AGE → months + LEAST(months_active, 12) for window cap + NULL handling.

Line 3: Karat Live Coding (SDE Track)

def validate_tax_fields(form):
    rules = {
        'income': lambda v: isinstance(v, (int, float)) and v >= 0,
        'ssn': lambda v: isinstance(v, str) and len(v) == 11 and v[3] == '-' and v[6] == '-',
        'filing_status': lambda v: v in {'single', 'mfj', 'mfs', 'hoh', 'qw'},
        'dependents': lambda v: isinstance(v, int) and 0 <= v <= 20,
    }
    errors = []
    for field, check in rules.items():
        if field not in form:
            errors.append(f'missing: {field}')
        elif not check(form[field]):
            errors.append(f'invalid: {field}={form[field]}')
    return errors

5-Day Sprint

FAQ

Need TurboTax background for Intuit DS?

Not required. Take-home uses tax terms (AGI, itemized deduction, refundable credit) — pre-skim these or you'll burn EDA time.

Karat language?

Python / Java / Go / JavaScript / Ruby. DS: ~95% Python. SDE: ~70% Python, ~25% Java.

Take-home pass criterion?

Technical correctness > business intuition > delivery. ~70% pass-through to onsite per community reports.

Result timeline?

Usually 7–14 days; take-home is reviewed by 1–2 graders + Karat performance.

Preparing Intuit OA / VO?

We were glad to help this cohort pass the Intuit Data Scientist OA. Many candidates told us the take-home challenge isn't code — it's "land a PM-convincing business conclusion in 4 hours" — something LC grinding never builds.

If you're prepping Intuit, TurboTax, QuickBooks, Mint, or Credit Karma DS / SDE OA / VO and feel directionless on business-case delivery, contact oavoservice. We tailor OA / VO assistance to your gaps.

👉 Add WeChat: Coding0201 — grab the Intuit prep pack.

Contact

Email: [email protected]
Telegram: @OAVOProxy