Traditional interview preparation is time-consuming, generic, and often lacks personalization. Candidates struggle to get targeted practice, actionable feedback, and domain-specific mock interviews tailored to their background, resume, and job description.

• Personalized Interview Simulation: The bot generates interview questions based on the user's resume, job description, and career goals.
• Domain-Specific Practice: Supports multiple domains (Engineering, Management, HR) with relevant question sets.
• Automated Feedback & Scoring: Provides instant feedback and scoring to help users identify strengths and weaknesses.
• No Resume/JD Required: Can function with just user profile information, making it accessible to all users.

The Interview Bot MVP is a Streamlit-based web application that simulates real interview scenarios. Users can upload their resume and job description (optional), fill in their profile, and select a domain for the interview. The bot generates questions, collects answers, and provides feedback and a summary at the end.

• Frontend/UI: Streamlit
• Backend/Logic: Python
• Vector Embeddings: sentence-transformers (MiniLM)
• Database: ChromaDB (for document and vector storage)
• PDF Parsing: pdfplumber
• Environment Management: python-dotenv
• Other: langchain, Google GenAI, pandas, numpy

[User] 
   |
   v
[Streamlit UI] <--> [Session State]
   |
   v
[InterviewerAgent] <--> [ChromaDB] <--> [SentenceTransformer]
   |
   v
[Feedback & Scoring]

• User interacts with the Streamlit UI.
• Session State manages user data, uploaded files, and interview progress.
• InterviewerAgent generates questions and feedback using user data and embeddings.
• ChromaDB stores user documents and their vector representations.
• SentenceTransformer encodes text for semantic search and question generation.

1. User enters profile info and (optionally) uploads resume/JD.
2. User selects a domain and starts the interview.
3. The bot generates questions using the InterviewerAgent, leveraging user data and embeddings.
4. User answers questions; responses are stored and scored.
5. At the end, the bot provides a summary and a score, with a full Q&A recap.

• Job seekers preparing for interviews in specific domains.
• Students practicing for campus placements.
• Professionals seeking targeted feedback for career transitions.
• Recruiters or trainers simulating interview scenarios for candidates.

• Add support for more domains and question templates.
• Integrate with external job boards and resume parsers.
• Enable voice-based Q&A and feedback.
• Advanced analytics and progress tracking.
• Multi-language support.
• Team/enterprise features for group training.

• Abhishek Kevin Gomes
• Ankur Singh