Visionary Engineer

Llama 2 70B Model Evaluation, Deployment, and Fine-Tuning

Aug 2024 - Present

Associated with Udacity

In this capstone project, I focused on evaluating, deploying, and fine-tuning the Llama 2 70B model. By using a specialized IT domain dataset, I transformed the pre-trained model into a domain expert. The project involved data preprocessing, model evaluation, and implementation of fine-tuning techniques to achieve optimal performance. The outcome was a highly accurate and efficient domain-specific language model ready for real-world applications.

Skills: Deep Learning, NLP, Model Fine-Tuning, Transfer Learning, AWS Sagemaker, and other related services

Medical Insurance Premium Prediction with Machine Learning

Aug 2023 - Aug 2023

This project involved mastering the theoretical foundations and practical use of neural networks through Keras and TensorFlow. I built, trained, and evaluated neural network models, gaining hands-on experience. I ensured data accuracy through meticulous cleaning, used feature engineering techniques, and visualization methods. I also extended the analysis by integrating linear regression for enhanced insights and predictions. This project developed my versatile skill set in machine learning, covering neural networks, linear regression, data preprocessing, and visualization.

Skills: Machine Learning, Artificial Neural Networks

Research Project: iVision HHID - Handwritten Hyperspectral Images Dataset for Forensic Analysis and Forgery Detection

May 2023 - Jul 2023

Collaborating with a team from IST University on the iVision HHID project, we focused on enhancing model accuracy for handwritten hyperspectral images in document forensic analysis. Through advanced hyperspectral imaging techniques, we collected diverse handwritten documents, encompassing various styles, inks, and papers. Employing cutting-edge image pre-processing methods like noise reduction and normalization, we optimized data quality. Meticulous annotation and ground truth generation ensured dataset reliability, covering attributes such as age, authorship, and ink analysis. Data augmentation techniques like rotation, translation, scaling, and synthetic noise injection enhanced dataset diversity. The resulting iVision HHID dataset serves as a benchmark for forgery detection, authorship verification, and ink analysis. This project demonstrates my expertise in image processing, data annotation, and collaborative research in hyperspectral imaging-based document analysis.

Skills: Spectral Analysis, Deep Learning, Machine Learning

Secure Multi-Client Chat Room Server

May 2023 - Jun 2023

Associated with National University of Computer and Emerging Sciences

This project implements a secure multi-client chat room server using Python's socket and SSL libraries. The server enables real-time text-based communication among multiple clients within a secure environment. It establishes a secure socket connection using SSL/TLS encryption, ensuring confidentiality and integrity. Clients connect to the server by providing the server's IP address and port number. Upon connection, clients receive a welcome message and are prompted to enter their name. The server handles each client connection in a separate thread, keeping track of connected clients and their addresses. The server broadcasts messages to all clients, with the sender's name as a prefix. When a client sends the "#quit" command, they are gracefully disconnected, and the server notifies all clients about the departure. This secure server facilitates seamless multi-client chat interactions, ensuring message delivery and maintaining the chat room's integrity through SSL/TLS encryption. It supports multiple simultaneous connections, providing a safe and user-friendly environment for communication and connection-building.

Skills: Python (Programming Language), Network Security, Virtualization, Infrastructure Solutions, Networking, Socket Programming, Troubleshooting, Communication Protocols, SDKs

Building a Linux Web Server on Raspberry Pi 4 with Chat Room Application

Apr 2023 - May 2023

Associated with National University of Computer and Emerging Sciences

Designed and built a Linux web server on a Raspberry Pi 4, along with a multi-client chat room application. Set up an optimized Linux environment, configured a web server (Apache2), and developed a user-friendly interface for real-time communication. Implemented security measures, including firewalls, encryption, and user authentication. Stored chat data using SQL. Deployed the project on the Raspberry Pi, enabling accessibility via its IP address. Demonstrated proficiency in network programming, Linux system administration, server deployment, and chat room application development.

Skills: Python (Programming Language), Network Security, Linux Server, Virtualization, Infrastructure Solutions, Networking, Apache2, Raspberry Pi, Debian, Communication Protocols, SDKs

Diabetic Retinopathy Identification Using Machine Learning

Apr 2023 - Jun 2023

Associated with National University of Computer and Emerging Sciences

This project focuses on utilizing machine learning models and transfer learning techniques for the accurate identification of diabetic retinopathy. The objective is to develop an efficient system for the early detection and diagnosis of this diabetes-related eye complication. A curated dataset of retinal images from Kaggle is used, accompanied by labels indicating the severity of retinopathy. Rigorous data preprocessing and annotation procedures were implemented to ensure dataset quality. Various machine learning models, including custom-designed convolutional neural networks (CNNs) and pre-trained models such as MobileNetV2, were trained and evaluated. Transfer learning techniques were employed to enhance model performance. The project showcases expertise in data collection, preprocessing, model design, and the utilization of state-of-the-art machine-learning techniques for medical diagnosis.

Skills: Deep Learning, TensorFlow, Transfer Learning, Machine Learning

Design and Implementation of Light Intensity Measurement using LDR and AVR Microcontroller

Oct 2022 - Dec 2022

Associated with National University of Computer and Emerging Sciences

During our project of Light Intensity Measurement using LDR and AVR Microcontroller, we gained hands-on experience in using MicroChip Studio for AVR programming and Proteus for simulation of the design. We also learned how to interface LDR with AVR microcontroller, how to use UART as a communication protocol for data transfer between microcontrollers and display the real-time data on a 16*2 LCD. Our project helped us understand the practical implementation of microcontroller-based systems and enhanced our skills in embedded system design.

Skills: C (Programming Language), Debuggers, Assembly Language, Leadership, Communication Protocols

Designed an Internetwork for our Campus (FAST University)

Dec 2022 - Dec 2022

Associated with National University of Computer and Emerging Sciences

Designed and implemented an internetwork using Cisco Packet Tracer to connect FAST University campuses and the Head Office. Created separate networks for Electrical Engineering (EE), Computer Science (CS), and Business Administration (BBA) departments at each campus with strategic IP addressing. Utilized routers for inter-campus connectivity and employed static routing or RIP based on the number of routers. Configured DHCP for IP and gateway allocation, ensuring efficient network management. Integrated servers at the Head Office for centralized services. This project demonstrated proficiency in network design, IP addressing, routing protocols, and DHCP configuration in establishing robust and interconnected internetwork for FAST University.

Skills: Internet Protocol Suite (TCP/IP), Networking, Troubleshooting, Leadership, Packet Tracer, Communication Protocols

Designed and simulated an FM system with SSB-FDM using MATLAB

Nov 2022 - Dec 2022

Associated with National University of Computer and Emerging Sciences

In Analog and Digital Communications, my partner and I designed and simulated an FM system with SSB-FDM using MATLAB. We gained practical experience in communication systems and a deeper understanding of modulation techniques. This project provided us with valuable skills in system design, simulation, and signal processing.

Skills: MATLAB

DC Regulated Power Supply

Nov 2021 - Dec 2021

Associated with National University of Computer and Emerging Sciences

Designed and Simulated 5V DC Power Supply with the help of Bridge rectifier, capacitors, and voltage regulator IC 7805 in Proteus. This project provided practical experience in designing power supplies, using simulation tools, and understanding electronic components and circuit design.

Skills: Proteus

Digital Communication System

Jun 2021 - Jul 2021

Associated with National University of Computer and Emerging Sciences

In this project, we have Simulate a simple communication system using two (8*2 mux) a parallel to serial and serial to parallel, communication is done via single line, simulating the design using the Proteus software for the designing purpose. This project gave us hands-on experience in digital communication systems and circuit design using simulation software.

Skills: Proteus

Designed university 3-D Computer Lab using AutoCAD

Nov 2019 - Dec 2019

Associated with National University of Computer and Emerging Sciences

In this semester project, I created a university 3-D lab design using AutoCAD tools and commands. This project improved my skills in 3D design and AutoCAD, providing practical experience in using design software for real-world applications.

Skills: AutoCAD

Smoke Sensor using Arduino

Oct 2019 - Nov 2019

Associated with National University of Computer and Emerging Sciences

Designed Arduino smoke level detector using MQ-2 sensor for measuring the level of smoke in the environment. This project enhanced my skills in using Arduino for sensor interfacing and real-time data processing. It provided practical experience in designing and implementing sensor-based systems for environmental monitoring.

Skills: C (Programming Language), C++