NLP Researcher

Education

Computer Science and Engineering, Bachelors

Work Experience

Student Researcher @ Center for Computational and Data Sciences Lab, IUB

• Worked on BaTECLaCor: A Novel Dataset For Bangla Text Error Classification and Correction (BLP Workshop at EMNLP2023)
• Finetuned LLMs for comparing with the transformer models I worked on
• Annotated dataset for my project and other lab projects related to computer vision

Publications

• BaTEClaCor: A Novel Dataset for Bangla Text Error Classification and Correction (BLP workshop at EMNLP2023)
  • Conducted research on error detection and correction in Bangla communication using machine learning and deep learning models
  • Performed ETL to curate, classify, and annotate a dataset with 10,000 real Bangla comments from Youtube
  • Experimented and demonstrated BanglaBERT’s accuracy in error-category classification
  • Experimented and highlighted BanglaT5’s effectiveness with a Rouge-L score of 0.845 for text correction
  • Contributed to the enchancement of linguistic precision in the Bangla-speaking community
• AI-based Software Testing (ISBM 2023)
  • Developed an in-depth review paper on AI-based software testing, showcasing benefits like efficiency enhancement, improved test coverage, and cost-effectiveness.
  • Highlighted AI technologies like machine learning and natural language processing to automate testing phases, including generation, execution and analysis.
  • Addressed critical challenges in AI-based testing such as data quality, interpretability, bias and model maintenance, offering actionable strategies tailored for small software firms
  • Key findings revealed enhanced testing efficiency,precise defect detection, and reduced costs through AI integration, alongside challenges like data quality and interpretability needing ongoing attention.
• Food Rotting Prediction and Detection System for WareHouse (IEEE Region 10 Symposium - TENSYMP2023)
  • Utilized real-time sensor data to train models and predict the likelihood of food entering the next stage of spoilage, triggering alerts as needed.
  • Programmed an Arduino UNO to read and interpret analogue inputs from multiple sensors simultaneously.
  • Applied logistic functions and coefficient estimation for predictive modelling and decision-making based on threshold probabilities.
  • Contributed to research aimed at enhancing food spoilage prediction and real-time monitoring systems.

Projects

• HealthCare App
  • Used Technolgies: Android Studio, Kotlin, Java
  • Description
    • An Android Application to manage health condition of patients
• Air Quality Monitoring System
  • Used Technologies: Django, Python, Pandas
  • Description
    • Built a webapp to visualize air quality of various regions of Bangladesh as per user query.
• CrimeChitro
  • Used Technologies: PHP, Javascript, Leaflet Js, MySQL
  • Description
    • Built a webapp that allows user to report crimes at a location in the map

Honors and Awards

• Best Undergraduate Student Researcher Award 2024
  • Issued by CSE Annual Research Day, Department of CSE, Independent University Bangladesh
• Champion in TechFest Autumn 2022 at Independent University,Bangladesh for the work titled “CrimeChitro” in the category of web-applications