Frequently asked questions

When you submit a query or program, the app sends your code, the question, and the generated schema or test cases to a large language model. The model reasons through your solution semantically rather than just string-matching, then returns a verdict, an explanation of what passed or failed, and where relevant a performance note and a best-practice tip. Because it grades meaning, a correct answer written differently from the reference solution still passes.

MySQL 8 is the default dialect for schemas, seed data, and reference solutions, including window functions and CTEs. Most answers written in standard ANSI SQL will also be accepted, and the mentor will point out where a construct is MySQL-specific.

SQL, Python, Java, PySpark, C and C++ (in one shared playground with a language toggle), plus a curated GCP Data Engineer question bank covering BigQuery, Dataflow, Pub/Sub, Composer, IAM, data modelling, warehousing and ETL/ELT.

Yes — every track, the AI mentor, the theory explanations, and the live AI interview are free to use with an account. There is no payment step and no card is required. Usage may be rate-limited during heavy load to keep the app responsive for everyone.

An account is required so your progress, solved library, and resume checkpoints can be saved and synced. You can sign up with email and password or with Google Sign-In.

Yes. Each track autosaves checkpoints of your active question, your editor buffer, and your session progress. When you return, a Resume prompt offers to drop you back exactly where you left off. Each subject has its own separate slot, so Python work never overwrites SQL work.

The interview runs in phases — introduction, technical depth, problem solving and wrap-up — with voice input and barge-in so you can interrupt naturally. Afterwards you get a scorecard rating technical accuracy, communication, problem-solving approach and depth, plus specific strengths and areas to work on. Scores are practice signals, not certifications.

Submissions are evaluated by AI reasoning against expected outputs and hidden test cases rather than being executed on a real server runtime. That makes feedback fast and lets it explain your logic, but it also means edge cases can occasionally be judged imperfectly — treat a verdict as strong guidance rather than a compiler result.

Yes. Each track has a Solved tab that stores every question you have completed together with your answer, the explanation, and the unique functions you used, so you can revise them before an interview.

Use the in-app feedback button, the contact page, or email pydikalvasrikanth@gmail.com directly. Please mention the track, the question you were on, and what you expected to happen — that makes fixes much faster.