ML-Jumbo
ML coding interview prep

Practice ML interviews —
and actually learn from every attempt.

Solve real ML interview problems and run them against hidden tests. Instead of a silent “wrong answer,” an AI judge explains your verdict, complexity, and what to improve — so every problem makes you better.

Free · sign in with GitHub, Google, or email.
🧩
Real ML problems
NumPy, PyTorch, ML fundamentals, DS&A, and system design — not generic puzzles.
🧠
Feedback that teaches
Every attempt gets a verdict, complexity, and concrete tips — so you learn why, not just whether it passed. No silent wrong answers.
▶️
Sandboxed execution
Your code runs against hidden tests in a secure sandbox, like a real interview judge.
📈
Progress tracking
Streaks, per-topic stats, and every past attempt in one dashboard.
🎤
Mock interview BETA
Talk through an open-ended ML question out loud in a timed mock, and get feedback on how clearly you communicated.
🎯
Personalized sets
Upload your resume and target jobs; get problems tuned to the skills they want.

ML-Jumbo

Loading…

00:00

Problem bank

Generate a fresh problem:
💡 Personalize is on, but your profile is empty — add target roles or a resume in the Personalization tab so there's something to tailor toward.
Title Topic Difficulty Solved

Progress

Activity (last weeks)

By topic

TopicAttemptedPass rateAvg score

Sessions

DateProblemVerdictScoreTestsTime

Collections

Personalization

Tune your practice toward the roles you're targeting. Optional, and off until you turn it on.

When on, your daily and generated problems lean toward your target skills. Turn it off anytime to return to the normal rotation.

💡 Add your target roles or upload a resume below so there's something to personalize from.

Target roles
Resume

Optional. Stored encrypted and used only to pick better problems — never shown to anyone. Delete anytime (removes the file and its analysis). PDF or DOCX, up to 2 MB.

Job-description sprint

Practicing for a specific job? Upload its description and pick an end date — problems lean toward that role until then. One sprint at a time; same encryption + delete as your resume.

🎤 Mock interview

A short, spoken mock interview — no coding. You'll talk through an open-ended ML question out loud and an AI interviewer will probe your thinking, then give you feedback on how clearly you communicated. About 10 minutes. BETA · 1 per day · needs microphone access.

Past mocks

10:00