About me

Hello I'm Purbasha (she/her), a graduate from the department of CSE, BUET. I am currently working as a lecturer in the department of CSE, BRAC University. My research interests lie in the fields of Systems and Networking, Distributed Systems, Human-Computer Interaction and Machine Learning. My current career goal is to pursue a PhD program in my field of interest.

For my undergraduate thesis, I worked under the guidance of Dr. Rezwana Reaz, focusing on the development of efficient fault-tolerant and elastic state management systems for Virtual Network Functions (VNFs) under various consistency requirements. We proposed a cluster based distributed state management system for network function that supports dynamic and efficient flow and state migration information between network functions.

I've recently worked with Salsabil Arabi and Adiba Shaira under the supervision of Dr. Tanzima Hashem on the development of a deep learning-based prediction model to infer the meetup POI of an individual at a particular time without the information of the companions associated with the particular meetup event. Currently, our manuscript is under review for publication at ACM TSAS.

Apart from my research, I am also fortunate enough to contribute to the academic community through teaching. I’m currently the instructor of undergraduate-level courses on Automata and Computability, Discrete Mathematics etc.

Resume

Research Interests

  1. Systems and Networking

  2. Distributed Systems

  3. Human-Computer Interaction

  4. Machine Learning

Work Experience

  1. Lecturer (Full-time)

    Department of Computer Science and Engineering,

    BRAC University

    June 2023 — Present

    Course Instructor:

    1. (Summer 2023, Fall 2023, Spring 2024, Summer 2024) CSE 331: Automata and Computability

    2. (Summer 2023, Fall 2023) CSE 230: Discrete Mathematics

    3. (Summer 2023) CSE 111: Programming Language II Sessional

Research Experience

  1. Cluster Based Distributed State Management of Network Function

    Systems and Networking, Distributed Systems

    Undergraduate thesis project under the guidance of Dr. Rezwana Reaz. We propose a distributed state management system for network functions (NFs) that ensures dynamic and efficient flow and state migration between NFs. Our approach introduces three innovative mechanisms to enhance the performance of stateful NFs. First, we implement multithreaded NFs to accelerate packet processing and optimize resource use. Second, we propose a cluster-based state replication scheme to minimize bandwidth consumption during flow migration, with states continually replicated across cluster members to facilitate smooth migration without overloading the network. Finally, we present an enhanced cluster-based load balancing algorithm to evenly distribute load and prevent bottlenecks. These approaches work together to maintain global state consistency without relying on a central authority. Experimental results demonstrate that our system achieves faster processing and reduced state migration overhead compared to existing solutions.

  2. Where Will You Meet on the Next Sunday?: A Deep Learning-based Approach for a Meetup POI Prediction

    Spatio-temporal Prediction, Deep Learning

    Recent work with Salsabil Arabi and Adiba Shaira under the supervision of Dr. Tanzima Hashem. In this paper, we introduce a deep learning-based model to predict an individual's point of interest (POI) for meetups at a specific time, even without knowing the companions involved. By leveraging transformer blocks, our approach incorporates features such as individual visit preferences, companions' visit patterns, and historical meetup data derived from check-ins and friendship networks. We demonstrate that these features enable highly accurate POI predictions. Validation on the US-based Foursquare dataset shows that our model significantly outperforms current state-of-the-art baseline models. Currently, our manuscript is under review for publication at ACM TSAS.

Education

  1. B.Sc. In Computer Science and Engineering

    Bangladesh University of Engineering and Technology

    April 2018 — May 2023

    CGPA: 3.68/4.00

    Notable Courses:

    CSE405- Computer Security CSE309- Compiler Design
    CSE453- High Performance Database Systems CSE321- Computer Networks
    CSE471- Machine Learning CSE313- Operating Systems
    CSE409- Computer Graphics CSE305- Computer Architecture
    CSE411- Simulation and Modeling CSE463- Introduction to Bioinformatics
    MATH245- Statistics and Probability MATH247- Linear Algebra

  2. Higher Secondary School Certificate (HSC)

    Sylhet Govt. Women’s College

    2015 — 2017

    CGPA: 5.00/5.00

    - Board Talent Pool Scholarship

  3. Secondary School Certificate (SSC)

    Blue Bird School and College, Sylhet

    2003 — 2015

    CGPA: 5.00/5.00

    - Board Talent Pool Scholarship

Technical Skills

  1. Programming Languages

    C/C++, Python, Java, Javascript, x86 Assembly, Bash

  2. Frameworks

    Docker, Django, NextJS, React, Tensorflow, Oracle DBMS, LaTeX

  3. Libraries

    Sklearn, Pandas, Matplotlib, Seaborn

  4. Dev Tools

    Visual Studio Code, Git

Projects

  • Gender Classification from Audio Clips

    ppml

    View on Github

    Tensorflow, Keras, Sklearn

    Term Project for CSE 472: Machine Learning Sessional offered by the CSE Department of BUET. In this project, we aim to build a machine learning model that can identify the gender of a person from their voice recording. In the process, we use two intermediary data representation format of the audio clips- Mel Spectrogram (Mel) and Mel-Frequency Cepstral Coefficients (MFCC).

  • RR-TCP: A Reordering-Robust TCP with DSACK

    ns3

    View on Github

    NS3

    In this project, Simulation and implementation of a congestion control algorithm on two different topologies using NS3 - a network simulator. This was the term project for the CCSE 322: Computer Networks course offered by the CSE Department of BUET.

  • Vectorized Convolutional Neural Network from Scratch

    ppml

    View on Github

    Numpy

    A vectorized version of a Convolutional Neural Network using only numpy without any deeplearning frameworks. Training and testing of the developed model is done on the NumtaDB: Bengali Handwritten Digits dataset. This project is part of the final assignment of CSE 472: Machine Learning Sessional offered by the CSE Department of BUET. The final score was tested on a hidden dataset.

  • EventFly: An Online Event Management Tool

    forage

    Video Demonstration View on Github

    ExpressJS, Kubernetes, Docker, NAT Streaming Server, MongoDB

    EventFly is an online event management platform where users can host and attend online or offline events based on the user preferences. This was the term Project for CSE 408: Software Development Sessional offered by the CSE Department of BUET.

  • Cycloan: An Online Cycle Sharing Platform

    forage

    View on YouTubeView on Github

    Django, Oracle SQL

    Cycloan is a ride sharing platform where users can put up their cycles for renting it to others and vice versa. The motive behind this project was to design a RDBMS based system on using only raw SQL queries. This was the term project for the CSE 216: Database Management System Sessional course offered by the CSE Department of BUET.

  • Compiler from scratch

    compiler

    View on Github

    Flex, Bison

    A compiler for subset of C language is made from scratch using flex and bison. This assignments are for the CSE 310: Compiler Sessional course offered by the CSE Department of BUET.

  • 8-bit MIPS Design and Pipelined Execution from Scratch

    mips

    View on Github

    Logisim - Java Platform (Version 2.7.1)

    Designed 8-bit MIPS from scratch and introduced pipelined execution using Logisim simulator. This assignments are for the CSE 306: Computer Architecture Sessional course offered by the CSE Department of BUET.

  • xv6 Memory Management

    forage

    View on Github

    xv6 Operating System

    Developed the paging framework of xv6 operating system, as well as implementing many other functionalities of the operating system.

  • Ray Tracing

    forage

    View on Github

    Graphics, OpenGL

    Followed by implementation of the raster based graphics pipeline, implemented ray tracing using OpenGL. This was the term project for the CSE 410: Computer Graphics Sessional course offered by the CSE Department of BUET.

Achievements

Honors and Awards

  1. Dean’s List Scholarship

    2019

    This scholarship is granted to undergraduate students for their academic excellence.

  2. Sylhet Board Talent Pool Scholarship (HSC)

    2017

  3. Sylhet Board Talent Pool Scholarship (SSC)

    2015

  4. Bangladesh Mathematical Olympiad

    1. 2nd Runner-up at National Mathematical Olympiad

      2017

    2. Participant at International Mathematical Olympiad Team Selection Camp

      2015, 2016

    3. Participant at National Mathematical Camp

      2014, 2016

    4. Participant at Special Mathematical Camp

      2015

Leadership Skills

  1. Organizer

    BUET CSE FEST 2022

    June 2022 - August 2022 [Visit Our Facebook Page]

    Coordinated several inter-university competitions such as hackathon, programming contest, deep learning competition, AI contest; as well as cultural programs on behalf of the graduating class.

  2. Student Tutor

    2018 - 2022

    Tutored many students ranging from middle school, high school and undergrad-level students during my undergrad years.