Spearheaded the development and production deployment of an email text intent classification model, meeting intent detection accuracy requirement of 90+%.Architected and deployed a self-healing prompt optimization pipeline, which autonomously adapts and fine-tunes prompts, reducing time to develop solutions by over 50% and increasing accuracy by 20-40%.

Performed pre-training, fine-tuning, evaluating, and delivering novel state-of-the-art solutions end-to-end in the generative ai space. This work leveraged large language models (decoder and encoder) and required significant research capabilities. Owned, Ideated, and Delivered key ablations enabling better Binary Embedding performance. Leveraged advanced distributed training and inferencing techniques reducing ablation time over 50%.Delivered inference improvements reducing latency by over 30%, enabling successful delivery of production models meeting product requirements.Regularly received top speaker scores at large external and internal AWS events (Re:Mars, Summits, and more.)Began transition to Senior Applied Scientist after successfully publishing findings in top-tier conferences, reflecting a strategic shift toward research-driven innovation and leadership within the organization.Led effort and built solution to train decoder and encoder LLMs end-to-end with Trainium and Inferentia, leveraging TorchXLA, transformers, and Optimum-Neuron, reducing training time and cost.

Identified distributed training opportunities for Apache MXNet and PyTorch customers on SageMaker culminating in a blog post on AWS, with the corresponding code in git (Python, MXNet, Docker, git). Train and deployed state-of-the-art defect detection solution on the edge for manufacturing use-cases, led development of novel search and ID resolution solutions.Delivered multiple high-impact solutions internally and externally. Impacting products like Amazon Forecast, Textract, Comprehend and Bedrock. Assisted multiple large AWS Customers with training and fine-tuning SOTA models in the areas of CV Keypoint detection, Diffusion, GPT, and other models from scratch.Co-chaired the distributed training leadership team, after successfully assisting multiple large customers in training large LLM and Diffusion Models. Published multiple workshops, tutorials, and papers for highly regarded internal AWS Conferences.Mentored many successful Junior Scientists and Interns resulting in high impact and follow-up internships.

● Full stack developer, increasing value to customer through numerous developments, leveraging Angular 4, Java, MongoDB, Centos, Docker, Neo4j, MATLAB.● Built a data science team and led them for a project focused on Anomaly detection in Time Series Telemetry data. ● Began a data science seminar, training computer scientists in Machine Learning and Data Science (Python3 with TensorFlow).● Technical contributor for a team working on state of the art Semi-Supervised Neural Networks for detecting objects in noisy images (Python3 with PyTorch).● Performed research on applying Neural Networks to anomalous signal detection in high noise environments (Python3 and MATLAB).● Created additional revenue streams through utilizing Machine Learning to pilot a program in determining component quality for testing.

● Updated legacy software for testing Hard Drives utilizing C, Python, and MATLAB while helping engineers with statistical problems that arose within different areas related to the future technology at Seagate.● Demonstrated an ability to learn independently at an accelerated rate with little prior knowledge in Engineering.● Assisted team member with trying to optimize servo arm placement using machine learning, leading to improved testing performance.

• Demonstrated an ability to independently learn in grading courses with no prior experience, showcasing a mastery of Undergraduate concepts and an ability to apply knowledge from familiar materials.• Displayed responsibility and integrity while working on a team with professors in creating exams, and grading while following guidelines from FERPA.

• Worked collaboratively with many different fields at NDSU, experiencing interdisciplinary group work while following FERPA guidelines, and practiced leadership in presenting material to students.• Developed software in C for research in kinetic theory of moving particles, acquiring advanced programming skills, enhanced independent learning skills, and practiced application of theory to computational code.

• Learned to adapt to different environments with fluctuating credit card laws and regulations.• Practiced teamwork with different branches and departments in a rapidly growing company while earning awards and bonuses for exceptional customer service and work quality.