elliott

Developed an information management system for Volvo using NestJS and React, optimized for AWS services (EC2, RDS, Auto Scaling) to enhance scalability and reliability. Designed and implemented a full-stack KPI management platform for Daimler, leveraging NestJS, React, and data visualization libraries for statistical analysis and performance metrics. Built a scalable ETL pipeline for Faurecia, automating real-time vehicle data migration from MySQL to Azure Data Lake and Cosmos DB using Azure Data Factory. Employed PySpark on Databricks to automate daily, monthly, and yearly reports, processing large-scale data efficiently. Utilized AWS S3 for intermediate storage and AWS Glue for data cataloging and transformation, improving data accessibility and performance.

Conducted deep learning research on CNNs, MLPs, RNNs, and LSTMs, focusing on their applications in lipreading technologies to improve speech recognition. Investigated the impact of liaison on mouth shapes, enhancing single-letter recognition precision and contributing to advancements in speech recognition accuracy. Developed and optimized neural network models to process and classify visual speech patterns, improving model efficiency and accuracy. Designed and implemented data preprocessing pipelines, utilizing OpenCV and NumPy, to enhance feature extraction for deep learning models. Synthesized research findings to inform the development of assistive communication devices, providing insights to improve interaction for individuals with hearing impairments.

Analyzed radar production data using pandas, NumPy, and scikit-learn, identifying key failure factors through exploratory data analysis (EDA). Developed and optimized machine learning models (Logistic Regression, SVM) to predict radar failures, improving failure detection accuracy by 2%.