Morgan Mastrangelo

Participated in developing an enterprise-scale multi-agent Generative AI platform, integrating GPT-4o, Claude, and Bedrock models with orchestrated reasoning workflows. Scaled Retrieval-Augmented Generation (RAG) systems using vector databases and semantic chunking strategies, reducing hallucination rates by 43% and boosting factual recall by 38%. Spearheaded the development of a model lifecycle m

Directed the end-to-end implementation of a cloud-native AI platform on GCP, leveraging BigQuery, Dataflow, Vertex AI, and Airflow for processing and serving over 10TB of data monthly. Developed full-stack AI solutions with FastAPI and GraphQL backends, embedding fine-tuned transformer models for tasks such as sentiment analysis and supply chain optimization. Engineered a distributed model deploym

Collaborated with Google Brain researchers on optimizing NLP transformer architectures for large-scale text classification, contributing to improvements that led to a 10% increase in model accuracy. Enhanced data preprocessing pipelines for massive unstructured datasets using Apache Beam and TensorFlow Datasets, reducing I/O bottlenecks and decreasing model preparation latency by 30%. Participated

Activities and societies: Dissertation: Advanced Techniques in Natural Language Processing and Large Language Models

Completed a Master's degree focusing on advanced computer science concepts. Specialized in Natural Language Processing and Large Language Models, culminating in a dissertation on these topics.

Obtained a Bachelor of Science in Computer Science, building a strong foundation in the field. This degree provided the essential knowledge for further specialization in advanced computing topics.