Chatbot-as-a-Service

Overview

Chatbot-as-a-Service is a production-oriented platform for deploying AI chatbots with trained NLP pipelines, caching layers, and REST APIs designed for real traffic.

Highlights

• Built a scalable AI-powered chatbot platform using FastAPI, NLP models, and AWS deployment.
• Achieved 76% intent classification accuracy using a trained NLP pipeline.
• Reduced inference cost by 30% through response caching and optimized API calls.
• Designed REST APIs handling 500+ requests/day with average latency below 200ms.

Architecture

The system separates ingestion, inference, and delivery. Intent classification runs through a dedicated NLP pipeline, while hot responses are cached to keep latency and cost low under load.

Role

Owned backend API design, model serving path, caching strategy, and cloud deployment. Focused on making the chatbot reliable enough for continuous production use — not just a demo notebook.