About

Joey Fitch is an experienced Machine Learning Engineer with over 6 years of relevant experience. Currently working at Performance Star LLC, he has developed an in-house Machine Learning platform from scratch in Java, which contains both deep learning and traditional modeling capabilities. His implementations specifically address client requirements of memory efficiency and outperform comparable models in Scikit-Learn and TensorFlow/Keras models. Joey is currently engineering a comprehensive Artificial Intelligence system to automate anomaly detection and process monitoring on semiconductor manufacturing systems. He is also building an adaptive platform for interpolating faulty sensors on semiconductor machines, which autonomously conducts feature engineering to determine predictors from over 1000 real-time input sensors using Dynamic PCA and pseudo-SHAP values. Joey has authored a novel machine learning algorithm to forecast time series from both sequential and non-sequential inputs, and he is currently working on publishing it. He has also constructed an LSTM architecture to monitor control signals over time, auditing dynamic physical relationships. Prior to his current role, Joey worked as a Data Scientist at Beeryou (raising series a), where he collaborated with the engineering team to construct BeerYou's database infrastructure in Google Firebase. He also worked as a Data Science Internship Leader at Verizon, where he led a team of 8 interns to cluster Verizon customers with similar web browsing preferences, used for targeted marketing. Joey holds a Master of Science (M.S.) degree in Statistics & Applied Math and a Bachelor of Arts (B.A.) degree in Mathematics from San Jose State University. He is well-versed in ML, Random Forest, CNN, Keras, Deep Learning, Scikit, AWS, Java, TensorFlow, infra, and Web technologies. Joey is constantly reading the latest statistical research papers to implement novel machine learning tools and improve performance against metrics of accuracy, speed, and memory efficiency.