Michael Montalbano
Development
New Jersey, United States
Skills
Machine Learning (ML)
About
Michael Montalbano's skills align with IT R&D Professionals (Information and Communication Technology). Michael Montalbano appears to be a low-to-mid level candidate, with 2 years of experience.
View more
Work Experience
Data Scientist
Lockheed Martin
January 2023 - Present
- Enhanced AEGIS Laser combat system capabilities by implementing network tracing to the cloud. Additionally, leveraged data science expertise as a Data Scientist for SENSIS, our warfare simulator product. KEY HIGHLIGHTS Developed a Multi-Agent Reinforcement Learning (MARL) AI system that significantly increased win rates in a military simulation (32%). Developed interactive workspaces demonstrating our data analysis techniques for government clients and inhouse modelers evaluating military scenarios. Spearheaded the implementation of containerized microservices using Kubernetes, enabling seamless communication between services. Additionally, led the integration of Kafka to achieve full system traceability, a standard now adopted across multiple teams. Developed modular software features using Java, SpringBoot, and C++ for a complex embedded systems, the AEGIS Laser Weapons System for Missile Defense.
Generative AI Research Associate
National Severe Storms Laboratory
January 2022 - December 2023
- Architected and implemented a generative AI pipeline for severe weather forecasting. KEY HIGHLIGHTS Developed a U-Net variant hail prediction model that outperforms existing models (CSI/POD/FAR = 0.33/.53/0.2) at 20 mm Leveraged time-delayed fields and multi-height radar data to optimize model accuracy. Through feature selection, identified a critical subset of 5 out of 25 variables for accurate hail prediction, contributing and confirming valuable scientific knowledge about hail-formation. Developed hail prediction model with U-Net variant. Performance on Hail Size: CSI/POD/FAR = 0.33/.53/0.2, a significant improvement over models of this kind. This hail prediction model, informed by key atmospheric features, empowers researchers to better assess storm intensity.
Education
The University of Oklahoma
Master's degree in Data Science
January 2019 - January 2021
South Dakota School of Technology
Master's degree in Physics
January 2016 - January 2019
