{"slug": "i-built-an-ai-interview-coach-for-my-friend-powered-by-open-weight-ai", "title": "I Built an AI Interview Coach for My Friend — Powered by Open-Weight AI", "summary": "A developer built Interview AI, a personalized interview-preparation web app for a friend preparing for software engineering interviews, using Next.js, TypeScript, and Google's open-weight Gemma model. The tool takes a candidate's background and target role as context, then generates tailored interview questions and evaluates their answers with AI feedback, with the model layer kept replaceable rather than tied to a closed provider. The project was submitted to the Hacktoberfest Weekend Challenge under the Best Use of Gemma category.", "body_md": "*This is a submission for the [Hacktoberfest Weekend Challenge: Build for a Friend](https://dev.to/challenges/hacktoberfest-weekend-2026-10-01)*\n\nI built **Interview AI**, a personalized interview-preparation assistant for a friend who was actively preparing for software engineering interviews.\n\nThe problem was simple: generic interview-preparation tools can generate questions, but they don't necessarily understand **the specific candidate, their projects, their experience, or the role they're applying for**.\n\nInterview AI turns that preparation into a personalized experience.\n\nA candidate can provide their background and the role they're preparing for, and the application generates relevant interview questions and helps them practice answers based on their actual experience.\n\nThe goal wasn't to build another general-purpose AI chatbot. I wanted to build something small that my friend could actually use before an interview.\n\n**Live application:** [https://interview-ai-azure-eight.vercel.app/](https://interview-ai-azure-eight.vercel.app/)\n\nThe application is designed around a simple loop:\n\n**Candidate context → Interview question → Candidate answer → AI feedback → Better preparation**\n\n**GitHub:** [Repository](https://github.com/Roushan-kr)\n\nThe project is built as a web application with a focus on keeping the interview workflow simple and usable rather than turning it into a generic AI chat interface.\n\nThe application is built with **Next.js, TypeScript, and an open-weight AI model** at its core.\n\nThe AI is responsible for understanding the candidate's context and generating interview interactions rather than simply returning predefined questions.\n\nThe basic architecture is:\n\n```\nCandidate\n    ↓\nNext.js Application\n    ↓\nInterview Context\n    ↓\nOpen-weight AI Model\n    ↓\nPersonalized Interview Question\n    ↓\nCandidate Answer\n    ↓\nAI Evaluation & Feedback\n```\n\nI intentionally kept the AI layer replaceable. The application isn't fundamentally tied to a single closed AI provider, which means the underlying model can be changed, experimented with, or run through different inference setups as the project evolves.\n\nThe most important part of the implementation is that the **open AI model isn't just an optional feature**. It is the component responsible for creating the personalized interview experience.\n\nFor an interview-preparation tool, the candidate's information can be highly personal: their resume, projects, previous experience, weaknesses, and the roles they're targeting.\n\nUsing an open model gives me significantly more control over that AI layer.\n\nInstead of treating the model as a black-box API that I cannot inspect or replace, I can choose the model, change the inference setup, experiment with prompts and behavior, and eventually run the system locally if the use case requires stronger privacy.\n\nThat matters especially for a tool built around someone's personal career information.\n\nOpen innovation also changes the economics of experimentation. I can prototype, evaluate different models, and potentially run inference without being permanently locked into a proprietary API or its pricing and availability.\n\nFor this project, **open AI means ownership and flexibility over the part of the product that actually makes it useful.**\n\nI am entering the **Best Use of Gemma** category because the project uses Google's open-weight Gemma model as the AI foundation for the interview experience.\n\nThe model is not being used simply as an add-on chatbot. It is responsible for generating and evaluating personalized interview interactions based on the candidate's context.\n\nBuilt for one person, but designed around a problem many candidates face:\n\n**How do I practice for *my* interview rather than just practice interviews in general?**", "url": "https://wpnews.pro/news/i-built-an-ai-interview-coach-for-my-friend-powered-by-open-weight-ai", "canonical_source": "https://dev.to/roushandev/i-built-an-ai-interview-coach-for-my-friend-powered-by-open-weight-ai-48gh", "published_at": "2026-10-05 17:46:43+00:00", "updated_at": "2026-10-05 17:47:46.236805+00:00", "lang": "en", "topics": ["ai-products", "ai-tools", "large-language-models", "generative-ai"], "entities": ["Interview AI", "Gemma", "Google", "Next.js", "TypeScript", "Hacktoberfest", "Vercel"], "also_reported_by": [], "alternates": {"html": "https://wpnews.pro/news/i-built-an-ai-interview-coach-for-my-friend-powered-by-open-weight-ai", "markdown": "https://wpnews.pro/news/i-built-an-ai-interview-coach-for-my-friend-powered-by-open-weight-ai.md", "text": "https://wpnews.pro/news/i-built-an-ai-interview-coach-for-my-friend-powered-by-open-weight-ai.txt", "jsonld": "https://wpnews.pro/news/i-built-an-ai-interview-coach-for-my-friend-powered-by-open-weight-ai.jsonld"}}