About

I am a Software Engineering researcher with a Ph.D. in Computer Science, specializing at the intersection of software engineering, natural language processing, and human-computer interaction. My research focuses on understanding and mitigating bias in software engineering communities, particularly through the development of AI-driven tools to detect and analyze discriminatory language in developer communications.

Currently, I am working as a Postdoctoral Researcher at Tulane University under the supervision of Hridesh Rajan, where I am exploring the role of large language models in content moderation, conversational dynamics, and trust in socio-technical systems. I am particularly interested in designing AI systems that are not only technically robust but also socially responsible and inclusive.

I completed my Ph.D. in Computer Science from Wayne State University, where I worked in the SEAL Lab under the supervision of Amiangshu Bosu, and earned my B.S. in Computer Science & Engineering from Bangladesh University of Engineering & Technology.

I pronounce my name as SY-muh sul-TAH-nuh.

Recent News

Education

  • Ph.D. in Computer Science
    Wayne State University  ·  Fall 2020 – Spring/Summer 2025
  • M.S. in Computer Science
    Wayne State University  ·  Fall 2020 – Fall 2023
  • B.Sc. in Computer Science and Engineering
    Bangladesh University of Engineering and Technology  ·  2012 – 2017

Publications

Journal & Conference Papers

Short Papers & Posters

Research Projects

Analyzing Toxicity and Gender Bias Impact on OSS Diversity

We are currently gathering GitHub dataset of pull requests and issue comments to explore the effects of toxic and gender-biased content on diversity in gender, ethnicity, and tenure. After that, using two state-of-the-art SE domain-specific toxicity and gender derogatory detection tools, we will automatically identify toxic and gender discriminatory texts and will train multivariate regression models, where three diversity indices measured using the Blau/Simpson index would be dependents, and the ratio of toxic /gender discriminatory texts would be one of the independents. To account for confounding factors, we will use various project characteristics such as the age of a project, the number of contributors, the number of total commits, the number of issues, the number of releases, the number of pull requests, and if the project has a code of conduct, as independents. We will assess the association between perceptible diversity and the prevalence of toxicity /gender-derogatory texts using regression coefficients and their significance (i.e., p-value <0.05).

MSR'24

Investigating LLMs for Detecting Sexist Content

We are assessing ChatGPT's capability to detect various forms of gender-based derogatory language. As an exploratory study, we collected a dataset from a previous study where the authors identified instances of gender-based maternal insults, commonly known as "mom jokes" and stereotypical remarks about women in an instant messaging chat environment. Utilizing the ChatGPT API, we sought to determine whether these comments could be accurately identified as mom jokes or stereotyping using both zero-shot and one-shot approaches. Our investigation has yielded encouraging findings regarding ChatGPT’s capacity to recognize sexist remarks. Specifically, we found that this model successfully identified 98.2% of texts containing maternal insults and 71.12% of texts containing stereotyping as positive instances. Hence, we can conclude that ChatGPT has demonstrated promising results in identifying specific forms of sexism.

PETRA'24 ESEM'25

Examining Gender Bias Among Software Industry Professionals

Based on the Social Identity Theory (SIT) framework, we selected four dimensions of biases for this study: bias in career development, bias in task selection, unwanted attention, and identity attacks. Through a vignette-based survey study, we investigated to determine the frequency of these biases, understand the association between demographic factors and being victims, explore the consequences of biases, and uncover the motivations behind biased behaviors. Our findings suggest that bias in career development and task selection are the two most prevalent among the four dimensions examined. More than two-thirds of the victims encounter biases multiple times. Demographic factors such as age, years of development experience, project experience, race, and location, in addition to gender, influence the likelihood of being a bias victim. Women are more than three times more likely than men to become victims of a bias. Additionally, workplace culture, along with stereotypical beliefs about women, often influences individuals to make biased decisions.

womEncourage'24 TOSEM (Under Review)

Automated Tool for Identifying SGID Content

Developed a rubric and annotated a dataset of 11,007 GitHub pull requests and issue comments to build a tool automatically identifying sexual and gender identity-based derogatory content. SGID4SE incorporates six preprocessing steps and ten state-of-the-art algorithms and implements six different strategies to improve the performance of the minority class. We empirically evaluated each strategy and identified an optimum configuration for each algorithm. In our ten-fold cross-validation-based evaluations, a BERT-based model boosts the best performance with 85.9% precision, 80.0% recall, and 82.9% F1-Score for the SGID class. This model achieves 95.7% accuracy and 80.4% Matthews Correlation Coefficient. Our dataset and tool establish a foundation for further research in this direction.

Rubric SGID4SE ESEM'21 ASE'22 TOSEM'25

Gender Bias in OSS Code Review

To identify whether the outcomes of participation in code reviews (or pull requests) are influenced by the gender of a developer, we conducted regression analysis on data of pull request approvals, review times, and reviewer participation from 1,010 GitHub and Gerrit projects divided into 14 datasets. Our results find significant gender biases during code acceptance and significant differences between men and women in terms of code review intervals, with women encountering longer delays than men in three cases and the opposite in seven. Moreover, our results indicate reviewer selection as one of the most gender-biased aspects, with 12 out of 14 datasets exhibiting bias. Since most of the review assignments are based on invitations, this result suggests possible affinity biases among the developers.

Replication ICSME'21 EMSE'23

Smart Home Device Use by Airbnb Hosts

An increasing number of Airbnb hosts are using smart home devices to manage their properties; as a result, Airbnb guests are expressing concerns about their privacy. To reconcile the tensions between hosts and guests, we interviewed 10 Airbnb hosts to understand what smart home devices they use, for what purposes, their concerns, and their unmet needs regarding smart home device usage. Overall, hosts used smart home devices to give remote access to their home to guests and safeguard their investment properties against misuse. They were less concerned about guest privacy and felt that smart home devices provided unique value to guests and, thus, a competitive advantage over other Airbnb properties.

Replication CHI'20

Experience

Professional Experience

  • Software Engineer Intern  ·  Meta, Menlo Park, CA, USA  ·  Summer 2024
  • Software Engineer  ·  Reve Systems, Dhaka, Bangladesh  ·  March 2017 – April 2019

Teaching Experience

  • CSC 4996 Senior Capstone Project  ·  Fall 2022, Winter 2024
  • CSC 4421 Computer Operating Systems Lab  ·  Winter 2022, Winter 2023, Fall 2023
  • CSC 2201 Computer Science 2 Lab  ·  Spring/Summer 2022
  • CSC 2200 Computer Science 2  ·  Spring/Summer 2021, Spring/Summer 2022
  • CSC 1101 Problem Solving & Programming Lab  ·  Fall 2020, Winter 2021, Spring/Summer 2021, Fall 2021, Spring/Summer 2023
  • CSC 1100 Problem Solving & Programming  ·  Spring/Summer 2023

News

Service

  • Reviewer, 42nd International Conference on Software Maintenance & Evaluation (ICSME 2026), Registered Report Track
  • Reviewer, 1st International Workshop on Software Engineering for and with Trustworthy LLMs (LLMTrust 2026)
  • Junior Program Chair, Mining Software Repositories (MSR) 2025, 2023
  • Reviewer, 47th International Conference on Software Engineering (ICSE) 2025 – Artifact Evaluation Track
  • Reviewer, Empirical Software Engineering Journal (EMSE), 2024
  • Student Volunteer, ACM Joint European Software Engineering Conference on Foundations of Software Engineering (ESEC/FSE) 2021, 2023
  • Student Volunteer, IEEE/ACM International Conference on Automated Software Engineering (ASE) 2022, 2021
  • Instructor of Machine Learning & Data Science courses, Community Engagement & K-12 Outreach Program, College of Engineering, Wayne State University, Summer 2021
  • Additional Reviewer, 44th International Conference on Software Engineering (ICSE) 2021 – Tool Demonstrations Track
  • Student Volunteer, 43rd International Conference in Software Engineering (ICSE) 2021
  • Reviewer, Website Design Competition for Middle and High School Girls, Michigan Council of Women in Technology (MCWT), 2020

Contact

Office
Paul Hall, Suite 518
New Orleans, LA 70118