Yashwanth Muthuramalingam Email & Phone Number

Frankfurt Rhine-Main Metropolitan Area

Stuttgart Area, Germany

Master's Thesis: "Image recognition in street scenes targeting object detection using Deep Learning" (Grade: 1.3)- Developed deep neural network ‘object detection’ models using Tensorflow framework to perform face detection behind vehicle glass and achieved an inference accuracy of about 95% (Faster R-CNN)- Increased the efficiency of face detection, significantly by about 30%, compared to the company’s previous detection model that used traditional computer vision techniques- Researched and evaluated various state-of-the-art convolutional object detection architectures to determine the best-fit model- Applied ‘transfer learning’ techniques to train the object detection models by employing pre-trained convolutional neural networks- Fine-tuned the fully-connected layers of Faster R-CNN model with random weight initialisation to output a resultant model with mean Average Precision (mAP) of 68.7% at IoU = 0.5- Prepared data by annotating class labels and bounding boxes for thousands of carefully selected images in varied environment settings, which turned out to be crucial for the model's success

Stuttgart Area, Germany

- Implemented geospatial data analysis and validation of Daimler’s supplier locations using HERE maps REST APIs.- Created Tableau dashboards for optimizing truckload capacities, aiding critical supply chain decisions.- Automated ETL processes and data quality analyses using Python with SQL server on the backend for periodic Tableau dashboard updates.

Master'S Degree, Data Science, 1.6

Bachelor'S Degree, Mechanical Engineering