Description

You will design and optimize speech recognition and natural language understanding systems for high-volume telephony environments.

This role is on-site.

Responsibilities

• Design and develop speech grammars for 80+ customer intents across 5 languages.
• Build and train intent classification and entity extraction models for accounts, amounts, dates, and transactions.
• Optimize speech recognition for 8kHz audio and noisy telephony conditions, maintaining 95%+ transcription accuracy.
• Implement context handling and dialogue management for multi-turn conversational interactions.
• Generate training datasets from failed interactions to continuously improve model accuracy and monitor call quality.

Required Skills

• 7+ years of experience in conversational AI, speech recognition, and NLU systems.
• Expertise in GRXML, ABNF grammar formats, and speech grammars.
• Proven track record managing multilingual platforms handling 10M+ annual interactions.
• Strong background in intent classification and entity extraction.
• Experience with NLU development and dialogue optimization.

Preferred Skills

• Acoustic model tuning and DTMF fallback mechanisms.
• Experience with multilingual language detection, dialect handling, and code-switching.
• Knowledge of sentiment analysis and A/B testing for conversational flows.

Key Skills

Education

Any Graduate