During the Machine Learning Fundamentals for Mechanical Engineering course at Georgia Tech, I collaborated on a project to develop and compare machine learning models for accurately predicting material stiffness based on microstructure features. This initiative was critical for tackling materials whose nonlinear behavior is time-consuming and computationally expensive to analyze through traditional methods.

The project involved building a comprehensive machine learning pipeline, starting with a dataset of 9,000 observations featuring the first 15 principal components of the microstructure data. My work included data validation, standardization (Z-score normalization), and training a suite of regression models to benchmark performance.

Key Contributions & Analysis:

• Model Implementation: Designed and implemented multiple regression models, including Linear Regression, Ridge/Lasso, Artificial Neural Network (ANN), and Extreme Gradient Boosting (XGBoost), to identify the optimal non-linear architecture.

• Hyperparameter Tuning: Conducted hyperparameter optimization (using methods like Random Search) on the ANN and XGBoost models to maximize generalization and minimize overfitting.

• Performance Evaluation: Performed detailed Root Mean Square Error (RMSE) analysis to compare model performance, demonstrating that simple linear models severely underfit the data (RMSE ~8.00).

Result & Impact:

• The Tuned Artificial Neural Network (ANN) was identified as the best-performing model, achieving a balanced and generalizable result with a Test RMSE of 2.32 (Train RMSE 1.91). This significantly outperformed the linear baseline and was less susceptible to overfitting than the XGBoost model (Test RMSE 2.65).

• This project showcases my ability to integrate computational modeling, advanced data science techniques, and engineering principles to create efficient, predictive tools for complex, non-linear problems in materials science.

During my time at Georgia Tech, I served as a core contributor to the ME 2110 robotics competition, where our team developed a high-performing mechanical system designed to execute complex, multi-faceted tasks under stringent constraints. 

As the lead CAD designer, I transformed initial concepts into precise, functional mechanical systems. My work directly influenced the robot's performance by introducing a dual-slider extension arm, which significantly improved deployment reliability, and a passive elastic depositor, which ensured compliance with strict size limitations while maximizing task efficiency. These innovations allowed our design to stand out in a highly competitive field.

To guide the design process, I collaborated on several engineering tools and methodologies, including the creation of a House of Quality (HoQ) to align customer requirements with engineering specifications. This analysis informed critical decisions, such as optimizing subsystem independence to allow simultaneous task execution. Additionally, I co-developed detailed function trees, morphological matrices, and evaluation matrices, which were instrumental in identifying the most effective design choices for the robot’s mechanical and operational subsystems.

In terms of manufacturing, I leveraged tools such as laser cutters, 3D printing, and rapid prototyping to fabricate and refine components. This hands-on work required careful material selection, including the use of lightweight but durable materials to balance cost, weight, and functionality. I also introduced modular designs, which allowed for faster repairs and adjustments during testing phases.

My contributions extended beyond technical skills to include team leadership and collaboration. I helped streamline workflows by creating clear timelines and deliverables, facilitating efficient communication, and ensuring that all design iterations aligned with our strategic objectives.

Through these efforts, our team achieved a highest scoring round of 106 points, ranking us competitively among other teams. More importantly, the project showcased my ability to apply engineering principles, adapt designs based on iterative analysis, and manage complex systems—all while meeting stringent requirements.

This experience not only strengthened my technical expertise in mechanical systems but also demonstrated my capability to lead innovative projects, utilize advanced design methodologies, and deliver tangible results. It reflects my readiness to tackle challenging engineering problems and contribute meaningfully to cutting-edge projects in professional environments.

At the Virginia Governor’s School, I designed and constructed a functional Arithmetic Logic Unit (ALU), a critical component in computer processors, using fundamental circuit design principles and hands-on prototyping. The project required combining theoretical knowledge of digital logic with practical implementation techniques.

Key Achievements:

• Digital Circuit Design:

  • Utilized logic gates, such as NAND, NOR, and XOR, to build a 4-bit ALU capable of performing basic arithmetic and logical operations like addition, subtraction, AND, OR, and bitwise negation.

  • Designed and tested the circuit on breadboards using components like quad NAND ICs, LEDs, and resistors for logic visualization.

• Schematic Development and Problem Solving:

  • Created comprehensive circuit diagrams to map out connections and ensure logical coherence.

  • Troubleshot and iteratively refined the design to resolve issues related to carry propagation, signal interference, and power efficiency.

• Hands-On Prototyping:

  • Assembled the ALU manually on breadboards, ensuring accurate wiring and reliable connections.

  • Integrated LEDs to provide real-time feedback for input, operation selection, and output verification.