I am an accomplished engineer with over a decade of experience in the automotive manufacturing industry. Since 2013, I have been a part of Denso, the world's second-largest supplier of automotive parts and systems. My journey began as a mold design engineering co-op during my junior year at Western Michigan University. Upon graduating in 2014, I quickly immersed myself in the complexities of Siemens NX design software while leading multiple mold builds.

Throughout my career, I have designed, refined, and implemented customer-driven modifications across hundreds of molds. My expertise spans finite element analysis (FEA) using Moldex3D, overseeing new mold developments, and conducting mold trials to ensure optimal performance. Leveraging simulation techniques, I have optimized molding parameters such as runner designs and gate locations. Additionally, I have successfully integrated additive manufacturing by designing dozens of conformal cooled 3D-printed steel inserts, resulting in significant cost savings in mold maintenance and cycle time.

In recent years, I have expanded my expertise into data science, recognizing its transformative impact across industries. At Denso, I compiled and analyzed 20 years’ worth of cooling fan warp data to identify how product design features influence warping behavior. Using predictive modeling, I have successfully forecasted warp characteristics in new cooling fans, minimizing tuning costs and driving substantial savings.

My passion lies in problem-solving and optimization. I thrive on tackling complex challenges, applying creativity, and developing innovative solutions that enhance efficiency and performance.

"We cannot solve our problems with the same thinking we used to create them." - Albert Einstein

Projects

❮ ❯

Overview: The ethical issues surrounding AI adoption in manufacturing were explored. Predictive Analytics and Maintenance, Business Intelligence, Robotics and Automation, Collaborative Robots (Cobots), Quality Control, and Optimization are all potential areas for AI to impact manufacturing. Job displacement is inevitable from new technology and will require companies to adopt AI in a resposible way. This presentation was for coursework at Eastern University. Check out my presentation on Youtube.

Skills: Ethics, Presentation

Overview: A convolutional neural network was created from scratch to further my knowledge on CNN architecture. The goal was to create a CNN that could predict whether a skin lesion is melanoma per the Kaggle competition. A dataset with over 30,000 images were used to create training, validation and testing datasets. Keras was used to create several models with different convolution, pooling, and dense layers. Parameters were tuned to find the optimized AUC-ROC score. Further improvement is needed to produce a more accurate model. Flask was used to present this model in a web application. Check out my project files on Github.

Skills: CNN, Flask, HTML, CSS

Overview: Cooling fan warp during the injection molding process was studied in effort to predict warp for new cooling fans. Traditionally, proto molds are built to validate the product design. This adds cost to the program and provides unreliable warp data for the production mold. I collected product design data, mold data, and measurement data on cooling fans in effort to build a machine learning model to predict warp based on the design. This approach eliminates the need for a proto mold and reduces the tuning cost of the production mold.

Skills: Machine Learning, Data Collection, Data Analysis

Overview: This notebook utilizes the mushroom dataset from The Audubon Society Field Guide to North American Mushrooms. The goal of this notebook is to display how a KNN Classifer can impute missing values and the importance of dimensionality reduction using PCA. Two simple models (Random Forest Classifier, Logisitc Regression) were trained and used to make predictions on whether a mushroom is edible or poisonous based on several attributes. The models were trained with and without PCA dimensionality reduction to compare. The performance measures: accuracy, precision, recall, training time were used to compare the effect of PCA on the models. Check out my Jupyter notebook on Github.

Skills: K-Nearest-Neighbor, PCA, Random Forest Classifier, Logisitc Regression