MockMate: a free interview coach for [friend's name] that runs on my own laptop
This is a submission for the Hacktoberfest Weekend Challenge: Build for a Friend What I Built Ragul is preparing for cloud and data analytics jobs in Tamil Nadu, and so am I, which means we both know the same problem: nobody to practise interviews with. Paid mock interview tools cost money, and practising on a friend gets awkward fast. MockMate asks interview questions on Python, Excel, SQL,…
This is a submission for the Hacktoberfest Weekend Challenge, aiming to build a free interview coach for a friend who is preparing for cloud and data analytics jobs in Tamil Nadu. Ragul is in the same boat, so he understands the issue of not having anyone to practice interviews with. Paid mock interview tools cost money, and practising with a friend gets uncomfortable quickly.
MockMate offers interview questions on Python, Excel, SQL, Power BI, and cloud basics. After you input your answer, it scores your response out of 10, pointing out what was good, what was missing, and providing a stronger sample answer. A progress tab tracks your average score per topic and highlights your weakest area.
The app runs on your own machine using an open-weight model, ensuring your answers stay on your laptop and that there's no cost involved. It was built for a friend as part of the Hacktoberfest Weekend Challenge, focusing on entry-level cloud and data roles in India. The code is concise, with an entire Python file of around 130 lines.
MockMate's key features include asking questions using the qwen2.5:7b model through Ollama, scoring answers out of 10, and storing every attempt in a local SQLite file. Streamlit provides the app's two-tab interface. The project was developed with the goal of addressing the vulnerability of interview practice, as responses are kept private on the user's own machine, there are no per-question charges, and the app functions offline once the model is downloaded.
Written by urgent.news from Dev.to's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.