AskHire is a small Streamlit-based chatbot that represents Vaibhav Karad and answers questions about his experience, skills, and projects.

Key features

• Chat UI built with Streamlit
• Uses Google Gemini (via google-generativeai) when an API key is provided
• Safe local fallback responder so the app remains usable without Gemini
• Tools for recording contact details and logging unanswered questions (integrates with Pushover if configured)
• Generates follow-up question suggestions to help users continue the conversation

Quickstart

1. Create and activate a virtual environment (recommended):

   On Windows (PowerShell):

   python -m venv .venv
   .\.venv\Scripts\Activate.ps1

   On Git Bash / WSL (bash):

   python -m venv .venv
   source .venv/bin/activate

2. Install dependencies:

   python -m pip install --upgrade pip
   python -m pip install -r requirement.txt

3. (Optional) Provide environment variables in a .env file or your environment. Supported variables:

   • GEMINI_API_KEY — API key for Google Generative AI (Gemini). If missing, the app falls back to a local responder.
   • PUSHOVER_TOKEN and PUSHOVER_USER — optional credentials to receive push notifications when tools are used.
   • GEMINI_API_KEY and PUSHOVER_* can also be provided via your environment or CI secrets.

4. Run the app:

   streamlit run main.py

Notes on behavior

• If GEMINI_API_KEY is set and google-generativeai is installed, the app will attempt to use Gemini and its function-calling features.
• If Gemini is not available (no package or no key), the app uses a lightweight local responder that:
  • Answers simple keyword-based queries from the candidate's summary/CV/LinkedIn
  • Detects contact information and triggers record_user_details
  • Logs unknown questions with record_unknown_question
• After each reply, the app generates suggested follow-up questions to guide the user. These suggestions are rendered in the UI (buttons) and can be used to prefill or auto-send follow-up messages.

Development notes

• Personal data files are read from the me/ folder (me/summary.txt, me/cv.pdf, me/linkedin.pdf). Keep sensitive data out of the repo or add me/ to .gitignore if needed.
• The code intentionally keeps the tool functions (recording contact or unknown questions) small and local — you can replace push() with other integrations (email, database, analytics).

Contributing

• Open an issue or submit a PR. Keep changes small and include tests or a short demo when changing behavior.

License

• This project is released under the MIT License — see LICENSE for details.

Contact

• For quick local testing, run the app and interact through the Streamlit UI at http://localhost:8501.