I'm Ayesha (she/her), a fresh graduate from the department of CSE, BUET. Currently, I am working as a full-time lecturer in the department of CSE, Canadian University of Bangladesh.
I am particularly interested to work in the fields of Natural Language Processing and Computer Vision. I am also interested in Computer security and Robotics. Since my undergraduate education, I have been actively engaged in academic research. I have worked on several interdisciplinary research projects in the field of ML, NLP and Computer Vision, both independently and in collaboration with professors. I also developed several Hardware projects using Arduino and ATmega32. I love working on real world problems that have an impact, and obsess myself till I find a solution. I am a regular participant in machine learning competitions.
In addition to my academic pursuits, I have a passion for music and can play three types of instruments. For recreation, I enjoy developing games, doodling, traveling, and watching sitcoms.
If you want to collaborate / want a pdf copy of my resume, knock me at : ayeshaathoi62 [at] gmail [dot] com
Undergraduate Thesis, NLP Group, CSE-BUET
We perform function call graph and code summary augmented few-shot prompting
for closed-source LLMs (GPT-3.5 Turbo, GPT-4, GPT-4o and Gemini-1.0-Pro), and also
fine-tune open-source LLMs (Llama 2, Code Llama, Llama 3) with quantized low-rank
adaptation to improve the state-of-the-art performance of
review comment and code refinement generation tasks, supported by a human evaluation study on CodeReviewer Dataset by Microsoft Research.
Manuscripting Under Preparation | Abstract
Ongoing Research Project, Field : Computer Vision
We generate a new image by combining the content of one image with the
style of another image while optimizing for minimal computational resource usage. We utilized Neural style transfer models : VGG16 and ResNet-50 and GAN model : CycleGAN.
With the help of these models, we are able to turn any picture into a painting of any famous artist. Paintings by Monet (Giverny in springtime), Pissarro
(Chestnut trees, Louveciennes, Spring - 1870), and Van Gogh(Starry Night) are used for result picture in the right side. Now, we are trying to transfer the style of calligraphy to a normal image.
We address complex task of object detection in driving environments across 9 districts in Bangladesh using BadODD dataset.
This dataset contains 9825 images with 78,943 Objects with 13 different classes.
We employed transfer learning using three different approaches, such as :
YOLO configurations, ranging from YOLOv6L6 to the advanced YOLOv8,
Novel transformer-based models such as rtDETR and CoDETR. We also explored Faster R-CNN.
It was a part of a competition, where we secured 8th position among around 100 teams.
Ongoing Research Project, Natural Language Processing X Computer Security
Working on extracting and analyzing attack behaviors from raw log data from the DARPA OPTC and DARPA TC E3 datasets. Due to current adverse situation, our work has been delayed.
We are using NLP techniques to extract attack behaviors from the raw log data.
Ongoing Research Project, Machine Learning
Using many types of regression models, we are trying to predict the ground level water in different regions of Bangladesh.
We are also using artificial neural network models to predict the ground level water.
For seasonal forecasting, we are using ARIMA, LSTM and GRU models.
It is a funded research project by WARPO (Water Resources Planning Organization) under the Ministry of Water Resources, Government of Bangladesh.
1. Implemented matrix transformation, eigen-value decomposition, image reconstruction using singular value decomposition.
2. Implemented adaboost algorithm with logistic regression on three classfication datasets with exploratory data analysis and preprocessing techniques.
3. Implemeted FNN from scratch | Relu, Softmax, tanh Activation | Dropout Layer | L1, L2, Binary Cross Entropy Loss | Report
4. Implemented PCA & clustering with EM algorithm on gaussian mixture models from scratch.
1. N-Puzzle Solver: Implemented A* heuristic (Hamming Distance, Manhattan Distance)
2. Latin Square Solver: Implemented csp value ordering (Forward Checking, Simple Backtracking) | Report
3. Exam Scheduling: Used local search (Largest Enrollment, Saturation Degree - DSatur Algo, Largest Degree, Random Ordering, Kempe-Chain, Pair-Swap) | Report
4. Minesweeper Solver: Utilized inference techniques (Implemented Random Moves, Safe moves, Discarded Diagonal Cells as neighbours)
1. Implemented 4-bit Arithmetic logic unit (ALU) with LED and ATmega32 desgining the soft circuit in Logisim | Flags : Carry, Sign, Overflow, Zero flags | Operation : Add, Add with carry, AND, Transter A, Increment A, XOR | Soft Circuit Diagram | Hardware Diagram
2. Implemented 32-bits Floating Point Arithmetic Unit (FPU) with ATmega32 and Logisim | 1 Sign bit, 10 exponent bits, 21 Fraction bits (Lowest bits) | The numbers are in normalized form. There are two flags U (underflow) and O (overflow) which are set if the result is too small or too large to be represented in the format | Report | Flowchart
3. Implemented a 4-bit MIPS Processor : Arithmetic Logic Unit (ALU), Control Unit and Datapath. For the simulation, we have used 1-RAM, 2-ROMs, 2-Shifters and 3-ALUs but in hardware implementation, we have used ATMega32 since we can program ATMega32 to add, subtract and store data | Report | Codes
1. Implemented assembly language programs including addition, subtraction, multiplication, division, and string operations.
2. Implemented microprocessor-based projects using Atmega32 :
a) Use of Polling for Digital IO.
b) Basic Bicolor LED matrix control with ATmega32: static, flash, and rotating display,
c) ADC and LCD module with ATmega32 | Youtube Video
3. Developed a Health Monitoring System using Arduino, pulse-oxymeter, DHT11 and GSM Module for covid patients.
4. Proteus Simulation for designing circuits with various tools.
Assembly and MicroController |
Proteus Simulation Code
Health Monitoring Project Code |
Report
1. Explored bash scripting
2. Implemented synchronization using pthreads :
Barber Shop, Master Worker, Read-Write Lock, Zemaphore
3. Implemented custom system calls : Trace & System Info to incorporate command line functionalities and lottery scheduling algorithm into existing xv6 pedagogical operating system.
4. Implemented copy-on-write paging mechanism for xv6 memory management.
5. Documented Git Commands.
All Code
1. Symbol Table : Stored info of occurance of entities : identifiers, objects, function names etc in CPP. | Code
2. Lexical Analysis : Tokenized the input source code interacting with symbol-table. | Code
3. Syntax and Semantic Analysis : Constructed bottom-up parse tree by shifting and reducing the tokens generated by lexer in Bison and Yacc. | Code
4. Intermediate Code Generation : Generated intermediate assembly code from the previous offlines outputs for the input source code. | ICG | ICG with Optimization
1. Implement public key cryptography system, AES, with Diffie Hellman and RSA key exchange algorithms interconnected via client-server socket.
2. Explored pedagogical malware (virus and worm) functionalities within docker containers and demonstrate buffer overflow attacks.
3. Experiment with firewall configurations
4. Implemented and demonstrate the features of MISP, an open-source Threat-Intelligence Tool | MISP Report
5. Implemented multi-modal Steganography and extracted hidden messages from stego objects | Steganography Code
1. setup Virtual Local Area Network (VLAN) using Cisco Packet Tracer and configure network components (Switch, Router, Trunk, NAT, PAT, ACL) | CheatSheet
2. We also implement a simple multi-threaded file web server using Java socket programming.
2. We then use Network Simulator 2 (NS2) to simulate wired and wireless network topologies with varying protocols across different network layers.
4. We experiment with WireShark to analyze network traffic.
5. Finally, we experiment with TCP Cubic-FIT simulation in NS2 and improved it || Report | Code
All Code
1. Ray Tracing: Implemented ray tracing for various shapes (cube, sphere, pyramids) with multiple perspectives and light sources, including a textured infinite checkerboard using a BMP file.
2. Raster-Based Pipelines: Developed a raster-based graphics pipeline in OpenGL covering transformations, clipping, scan conversion, and Z-buffer.
3. OpenGL Projects: Created projects such as an analog clock, transforming an octahedron into a sphere, and animating sphere and cube motions.
We developed a mobile application for parking management. The app allows users to book parking slots, pay for parking, and view parking history with various searching facilities. The app also includes a feature for parking lot owners to manage their parking lots and view parking history.
We designed BPMN, Mock-UI, ERD, Class, Collaboration, Sequence and State Diagram. Later, we developed a one module of the project. We have deployed our Database & Server in Render.
Diagram |
Module Code
An Online School inspired by Coursera as Covid hit the world. Here, 3 types of users can be created : Student, Teacher & Admin. The student can enroll themselves in different courses, test themselves through quizes. Teacher can create courses, add lessons, create quizes. Admin can manage all the users, courses, lessons, quizes.
Code
ERD
An Online platform where a patient can book appointment with doctors, order ambulances and check health by nurses.
Chatting service with doctors or nurse is also included. History will be saved and patient rate the services. And a hospital can see the rating and appoint nurse for the checkup. Online transaction facility is also included.
Code
Mock UI (Full Demo Look)
CGPA : 3.63/4.00
Last 2 year CGPA : 3.86/4.00
Undergraduate Lab Courseworks : Webpage
- Dean’s list award and university merit scholarship recipient in levels 3 and 4
Notable Courses : Machine Learning, Simulation and Modeling, Graph Theory, Artificial Intelligence, Bioinformatics, Software Engineering, Information System Design, Computer Security, Operating Systems, Compiler, Computer Networks, Data Structures and Algorithms, Database Systems, Computer Graphics, Discrete Mathematics, Object Oriented Programming
GPA: 5.00/5.00
Group : Science
Talent Pool Scholarship, 5th place holder (1st among girls) in Comilla board
Awards : Best in Academics award, Best Journal award, 1st Runner-up in Intra College Math Olympiad
GPA: 5.00/5.00
Group : Science
Talent Pool Scholarship, 8th position holder in Comilla board
Activities : Music, Chess, Handball, Band Party Drummer, Current Affairs Display Performer, Class Captain
Languages | C, C++, Java, JavaScript, Python, SQL, Assembly 8086, Bash, Latex, HTML, CSS |
Database | MySQL, OracleDB, PostgresDB |
Frameworks | PyTorch, Keras, Tensorflow, SpringBoot, Node.JS, React.Js, BootStrap |
Libraries | NumPy, Keras, Matplotlib, OpenCV, Pandas, Scikit Learn etc. |
API | REST, OpenGL |
Platform | Windows, Ubuntu |
User Research | Interview, Survey, Field study |
Research Methods | Mixed-method, Experiment, Data Scraping |
Communication Medium | English (Professional), Bengali (Native), Hindi (Professional), German (Intermediate) |
Tools/Software | Git, GitHub, Docker, render, IntelliJ, PyCharm, Codeblocks, Navicat, Visual Studio, WireShark, emu 8086, Jupyter Notebook, NS2, Google Colab, Kaggle, Logisim, Proteus |
Joined at CSE department under School of Science & Engineering. I find it a noble profession as it gives an opportunity to shape the life of others and pursue one’s passion for a lifetime. I love to interact and connect with my students.
Courses Teaching, Summer 2024 :Contributed to developing an Android App for Client-side verification of National ID Card images, using Android Studio, openCV and TensorflowLite.
IEEE Computer Society BUET Student Branch Chapter [2021-2023]
Publicity Committee Co-ordinator : 2022-2023 ; Media Committee Executive : 2021-2022
- Helped organizing competitions, career talks, research seminars and software development workshops
- Check out all of our events here.
Online Class Recording Management [2020-2024]
- maintained all recordings for my classmates in my youtube channel from the covid.
- Check out my Youtube Channel, although the playlists are unlisted.
- Tutored many students ranging from high-school to college-level
during my undergrad years.
ISSB Long Course-80 [2018]
- Yellow Card Holder, selected for Bangladesh Army Commissioned Officer Rank
- Medical Test : Passed
4 Day Long Test :
Day 1 : Intelligence Test, Picture Perception and Description Test, Personality Test, Essay Writing (Bangla, English), Medical Test
Day 2 : Group Discussion 1, Progressive Group Task, Half Group Task, Extempore Speech, Physical Ability Test, Interview
Day 3 : Planning Exercise (Written & Discussion), Command Tasks, Mutual Assessment, Interview
Day 4 : Assessment conference, Announcement of final results and Final Briefing
I love to develop games as I like to play games. I have developed some games using CPP, Pygame & JavaFX.
I sing Tagore Song. I have developed a great passion for music from childhood. I can play harmonium, keyboard and kalimba. I have performed in various cultural programs. You can visit my music playlist for more.
Apart from music, I love writing journals and stories. One of my Bangla journals got prize in 2017 on the occation of "International Mother Language Day".
I also love doodling, drawing, photoshop & adobe illustration.