Develop AI solutions for educational field● Customized generative/LLM based writing assistants and tools● Award winning AI scoring systems deployed nation-wide

● Providing deep learning solutions to a team of Geo-spatial scientists for drone imagery tasks, such as component inspection and damage detection.● Models deployed across various industries, including electric grid, solar farms, etc. ● Designing and managing cloud-based, distributed model training pipelines (AWS, CloudFormation)● Developing custom algorithms for VOT (video object tracking), image alignment, other in-house problems that arise.

● Develop deep learning models (in PyTorch, TensorFlow) for inference/prediction on time series data from wearable motion sensors● Deploy deep learning models to AWS cloud (Sagemaker, Lambda)● Mine historic motion sensor datasets for patterns, relationships that lead to actionable insights for customers

Responsible for research, development of all automated text scoring products (AES: automated essay scoring, SAS: short answer scoring), deployed in hundreds of school systems nationwide, covering over 30 states, and deployed as the backend text scoring engine for other big name vendors in educational testing such as ETS● Converted our automated scoring systems to deep learning systems (LSTM, BERT Transformer), reducing model training time from several hours to several minutes, while also improving accuracy.● Created "best-in-field" automated scoring systems for both essays and short-answers (see: Honors and Awards, below)● Developed state-of-the-art real time spell correction system (using Google N-gram database) now deployed in all automated scoring engines and in all MI's online writing sites.● Developed a new optimization algorithm to directly maximize novel objective function (Quadratic Weighted Kappa) most commonly used in automated test scoring (paper in preparation).