

Contributed to the implementation of a new search engine architecture for AI-powered chatbot shopping assistant incorporating RAG pipeline, vector storage, and LLM integration, significantly improving recommendation relevance and reducing customer search time.
Optimized AI-chatbot performance by testing workflows and designing enhanced prompts, resulting in improved inference accuracy, strengthened security, and increased response speed.
Implemented "LLM as a judge" approach to evaluate AI responses.
Participated in developing a web application that streamlined the creation of new microservices in the bank's infrastructure, boosting coding efficiency by approximately 30%.
Expanded production testing app with 2 critical user flows (notifications, async transactions), detecting errors in new service versions before rollout to other dev teams, cutting production incidents by ~10%.
Improved backend microservice for digital banking app (main page) by increasing unit test coverage from 60% to 90%, reducing errors in new feature deployments.