At MyTrek AI, I contributed as one of two core developers building MyTrek, an AI-powered learning assistant for personalized education. I developed backend and AI infrastructure, including RAG pipelines, chatbot memory, and conversational workflows.
At Ceres Technology, I built LightGBM models to estimate supply-chain delay risk and developed an analytics chatbot that let users query data and generate visualizations. I also created a dataset recommendation service that compared user data against more than 300k datasets.
At Digikala, I improved a trend-finding algorithm with parallel processing and built a recommendation system that increased add-to-cart rate by 70% and daily GMV by 202%. Earlier, at Part AI Research Center, I trained and evaluated speech models and led a team of four data scientists.

