NLP
Flunkey
NLP-Driven F&B Chatbot
Deployed intelligent chatbot using historical data with NLTK for intent classification and entity extraction. Enhanced contextual understanding by fine-tuning BERT and integrating Word2Vec embeddings, improving conversational relevance by 35% (F1-score). Built end-to-end TensorFlow pipeline achieving <500ms latency per interaction.
4 months
Project Duration
35% conversational relevance improvement
Key Achievement
5+
Technologies Used
The Challenge
Flunkey needed an intelligent chatbot for F&B digitization that could understand context, handle complex queries, and provide relevant responses with minimal latency.
The Solution
Deployed an NLP-driven chatbot with BERT fine-tuning for intent classification, Word2Vec embeddings for semantic understanding, and TensorFlow backend for real-time processing.
Results & Impact
- Improved conversational relevance by 35%
- Achieved sub-500ms response latency
- Handled 10,000+ customer interactions daily
- Reduced customer service costs by 50%
Project Metrics
35% conversational relevance improvement
<500ms latency
Technologies Used
NLTKBERTWord2VecTensorFlowIntent Classification
Project Details
Company
Flunkey
Duration
4 months
Team Size
2 engineers + 1 UX designer
Category
NLP