Sajeda Patel

Developed a dynamic stock trading simulation web application that immerses users in a competitive trading environment using real-time NYSE stock prices. Players receive an initial virtual cash balance and strategize to maximize their portfolio value by buying and selling stocks. The application features distinct player and admin portals, enabling an engaging and well-managed simulation experience.
Player Features:

● Real-time stock trading with live NYSE data.
● Portfolio tracking with current holdings, cash balance, and performance insights.
● Competitive leaderboard to track player rankings.

Admin Features:

● Configure games with custom start/end dates and trading rules.
● Player registration and game monitoring.
● Dashboard to oversee simulation performance and declare winners.

Key Contributions & Technologies:

● Designed and implemented buy/sell stock functionality using Node.js and integrated real-time NYSE stock prices.
● Created a responsive Graphical User Interface (GUI) using HTML, CSS, and JavaScript for both player and admin portals.
● Developed a robust back-end using Express.js and MongoDB for session and data management.
● Implemented unit tests to verify core functionalities like player registration and trading actions.

Team Collaboration:

● Worked in a Scrum team, actively contributing to sprints, updating the Kanban board, and participating in retrospectives.
● Collaborated with teammates on system architecture, UML diagrams, and sprint reviews.

Tools & Technologies:

● Languages/Frameworks: JavaScript (Node.js, Express.js), HTML, CSS, MongoDB.
● Testing: Created and executed unit tests using Node.js.
● Documentation: Prepared comprehensive README, sprint retrospectives, and process models.

Outcome & Future Plans:
Successfully implemented a fully functional stock trading simulation with scalable features. Future plans include enhancing user profiles, automating winner declaration, and expanding admin capabilities for better game management.

Developed a simulation-based autonomous drone control system using PX4, ROS 2 Humble, MAVROS, and Gazebo, focusing on building a complete software pipeline for drone communication, control, and real-time data interaction. The project aimed to enable programmatic control of a drone in OFFBOARD mode while understanding system-level communication between flight controllers and robotics middleware.

Key Contributions & Technologies:

● Built and deployed a PX4 SITL (Software-In-The-Loop) simulation integrated with Gazebo (gz_x500 quadrotor model) for realistic drone testing.
● Established communication between PX4 and ROS 2 using MAVROS and MAVLink protocol, enabling real-time telemetry and command exchange.
● Developed a control pipeline using ROS 2 topics and services to interact with drone states, including IMU, GPS, battery, and local position data.
● Implemented and tested OFFBOARD control logic, allowing programmatic interaction with the drone through external commands.
● Performed system-level debugging using PX4 logs to analyze issues such as EKF state estimation failures and preflight checks.

Research & Development Approach:

● Focused on understanding embedded flight control systems and middleware communication in robotics environments.
● Investigated real-world constraints such as sensor reliability and state estimation in autonomous systems.
● Iteratively debugged and validated system integration across simulation layers (PX4, MAVROS, ROS 2, Gazebo).

Tools & Technologies:

● ROS 2 Humble, PX4 SITL, Gazebo
● MAVROS, MAVLink Protocol
● Python, Linux (Ubuntu 22.04)
● Robotics Simulation & Autonomous Systems

Outcome & Future Plans:

Successfully developed a functional drone simulation pipeline with real-time communication and control capabilities. Future work includes achieving stable OFFBOARD flight control, integrating real drone hardware, and extending the system for autonomous navigation and mission planning.

Developed an intelligent medical image classification system to assist in pneumonia detection using chest X-ray images, focusing on building a reliable and generalizable deep learning solution for real-world healthcare applications. The project was developed through multiple iterative cycles, addressing practical challenges such as class imbalance, overfitting, and inconsistent model predictions.

Key Contributions & Technologies:

● Evaluated deep learning architectures including DenseNet121, ResNet50, and EfficientNet-B3, and refined the approach based on performance and generalization challenges.
● Designed and implemented a custom CNN model, applying data augmentation, dropout regularization, L2 weight decay, and class weighting to improve model performance.
● Applied early stopping and optimized prediction thresholds using F1-score analysis to ensure balanced and reliable classification.
● Achieved 88.2% accuracy, 90.4% precision, and 90.3% recall, outperforming initial benchmark models.
● Integrated the model into a Streamlit-based web application for real-time image upload and prediction with confidence scores.

Research & Development Approach:

● Conducted iterative experimentation to address model instability and improve generalization.
● Identified limitations in transfer learning approaches and refined the architecture accordingly.
● Focused on developing a practical and deployable AI solution, rather than only optimizing theoretical performance.

Tools & Technologies:

● Python, TensorFlow, Keras
● NumPy, Pandas, Matplotlib, Scikit-learn
● Deep Learning, CNN, Medical Image Processing
● Streamlit (Deployment)

Outcome & Future Plans:

Successfully developed a reliable pneumonia detection system with practical deployment capabilities. Future work includes improving model interpretability, integrating attention mechanisms, and extending the system for multi-disease classification.