Telegram bot for automatic interview (voice → AI analysis)
Developed a Telegram bot that fully automates the initial selection of candidates without employer involvement. The candidate starts the bot with the command /start, sequentially responds with voice messages to a series of questions, and the system transcribes the answers and generates a structured AI report for the employer.
What has been implemented:
- Step-by-step interview scenario: the bot asks questions in turn, accepts answers only in the form of voice messages, and automatically moves to the next question.
- Speech recognition: each voice message is transcribed into text using Whisper.
- AI analysis: after completion, the entire dialogue goes to GPT, which generates a report — a brief summary of the candidate, strengths and weaknesses, assessment of professional and communication levels, detection of template/AI-generated responses, and a final recommendation (recommend / consider further / do not recommend).
- The completed report is automatically sent to the employer via Telegram.
- All voice messages, transcriptions, and reports are saved.
- The list of questions is placed in a separate file — the client can change the questions themselves without code modifications.
- The solution is lightweight: it operates without heavy infrastructure and does not require constant maintenance.
Stack: Python, aiogram, Whisper (speech recognition), GPT (analysis and report generation).
Result: the employer receives a ready-made structured report for each candidate instead of hours of manual listening — the initial selection is fully automated, 24/7. A working demo is available upon request.
#TelegramBot #Python #aiogram #OpenAI #Whisper #GPT #ArtificialIntelligence #AI #Automation #Chatbot #BotDevelopment #HRTech #VoiceBot #AIAnalysis
What has been implemented:
- Step-by-step interview scenario: the bot asks questions in turn, accepts answers only in the form of voice messages, and automatically moves to the next question.
- Speech recognition: each voice message is transcribed into text using Whisper.
- AI analysis: after completion, the entire dialogue goes to GPT, which generates a report — a brief summary of the candidate, strengths and weaknesses, assessment of professional and communication levels, detection of template/AI-generated responses, and a final recommendation (recommend / consider further / do not recommend).
- The completed report is automatically sent to the employer via Telegram.
- All voice messages, transcriptions, and reports are saved.
- The list of questions is placed in a separate file — the client can change the questions themselves without code modifications.
- The solution is lightweight: it operates without heavy infrastructure and does not require constant maintenance.
Stack: Python, aiogram, Whisper (speech recognition), GPT (analysis and report generation).
Result: the employer receives a ready-made structured report for each candidate instead of hours of manual listening — the initial selection is fully automated, 24/7. A working demo is available upon request.
#TelegramBot #Python #aiogram #OpenAI #Whisper #GPT #ArtificialIntelligence #AI #Automation #Chatbot #BotDevelopment #HRTech #VoiceBot #AIAnalysis
Kharkiv 
Available for hire
Offer work