My main research focuses on neural information retrieval rankers, with a
particular emphasis on applications for federated search engines and datasets with limited resources.
I dedicate my efforts to analyzing and enhancing transformer-based document rankers to use them in
real-world applications. Following this path, I designed and implemented an app selection engine that
leverages these advanced rankers. Currently, I'm designing cross-market question-answering and
multi-resource systems and exploring the potential of automatic labeling for product question
answering, using the capabilities of OpenAI and open-source Large Language Models (LLMs).
I worked on different aspects of Information Retrieval and Natural Language
Processing under the supervision of Dr.
Saeedeh Momtazi. My research included improving a collaborative recommender system by using
different user profiling, also developing an expert finding system using graph embeddings and text
representation profiling, proposing a new model to represent users according to their past questions
and answers.
Software Engineer - Adanic (2017 -
2020)
I worked as a Java Developer on a Channel Manager System named Kariz, which
uses Java Spring Integration. This system makes an easy and effective management procedure of various
channels, and it provides a uniform layer of all services. I also designed a monitoring system using
elastic search, logstash, and kibana.
I was assigned to design an Investment Company
Portfolio Management System as my summer internship project. I
used Python to carry out some statistical analysis and implement the "Efficient Frontier" algorithm.
