×
This AI-powered chatbot provides an interactive platform to explore my resume, answering questions about my skills, experience, and qualifications.
For inquiries beyond the resume's scope, it transitions to a lightweight AI language model.
It is hosted on AWS Lambda Layers, a serverless environment with a strict 250 MB deployment limit for code and dependencies, which typically restricts the use of large language models.
To meet these constraints, I fine-tuned a compact T5-efficient-tiny model on an English OpenAssistant dataset and converted it to ONNX format for efficient inference.
Intent classification is handled by a quantized Sentence-Transformers MiniLM model, also converted to ONNX, balancing accuracy with size and speed.
This architecture enables running a conversational AI chatbot on a shoestring budget without relying on expensive hosting services.
The solution demonstrates my ability to apply advanced AI techniques with practical, cost-effective deployment strategies —delivering a scalable, fast, and user-friendly chatbot.