Related skills: Java, LLMs, NLP, RAG, Python, Azure, Docker, IntelliJ

• Developed Amelia, a full-stack AI-driven automation software, utilizing the Spring framework in

Java to enhance system efficiency and streamline workflows.

• Contributed significantly to the development and maintenance of Amelia’s RAG

(Retrieval-Augmented Generation) solution, Amelia Answers, by integrating an external

Python-based LangChain service for enhanced performance and functionality.

• Owned the creation and maintenance of the CQA (dynamic clarifying questions) portion of

Amelia Answers, including all prompt engineering.

• Fine-tuned Phi-3 and Llama-3 models, including synthetic data generation and training, for intent

and entity extraction, to achieve 98% performance on internal entity and intent classification

datasets.

• Built and maintained a Python service that returns tabular data in conversational format using

LLMs, optimizing user request fulfillment through advanced prompt engineering.

• Built and deployed a MS365 CoPilot message extension plugin in Typescript to integrate Amelia into MS365 CoPilot.