AI Fellow for Executive Enablement · Ibotta
Building explainable AI solutions that free team members up for higher-impact work.
AI · software · research
I’m an AI engineer with a history of AI enablement, software engineering, and research, graduating from Cal Poly San Luis Obispo with both a bachelor’s and master’s in computer science. I like turning difficult technical product problems into products that people can actually use.

A history of building, researching, and leading across applied AI and software engineering.
Building explainable AI solutions that free team members up for higher-impact work.
Training adversarial models against network intrusion detection systems and studying resilient NDR solutions.
Lead the CS+AI community at Cal Poly by organizing technical projects, workshops, hackathons, and events; supporting members; and connecting students with opportunities across computer science and artificial intelligence.
A fellowship in partnership with CodePath and Anthropic, serving as a precursor to the Claude Core program.
Built a Google Workspace RAG chatbot, Vertex AI search infrastructure, and Terraform-based CI/CD.
Led eight developers building a full-stack volunteer management platform used by nonprofit teams.
Built an AI research assistant and a 20,000-node knowledge graph for the San Luis Obispo City Council.
Worked with Franz Kurfess and a research team to build a deep-learning search-and-rescue platform that used machine learning to support search operations and provide decision support.
Technical lessons, personal hot takes, life lessons, and other random thoughts.
A few systems, experiments, and collaborations that represent how I work.
Syndicated data from internal databases and the web into a daily summary of postings for internship seekers.
TypeScript · Node.js · Discord API · Notion APIWorked on a five-person data pipeline that visualized user sentiment across subreddits.
Python · K-means · MongoDB · Text embeddings · Similarity searchCompared Genesis and Isaac Gym across local devices, AWS EC2, and Brev, evaluating five simulations at up to 200,000 environments simultaneously.
Python · Genesis · Isaac Gym · Brev · EC2 · HTTPerf · JMeterImplemented Unix file-block operations and a TinyFS file system with permissions, directories, timestamps, fragmentation, and renaming.
C/C++ · Multithreading · Vim · OS · HDD · I/ODeveloped a multi-user client-server chat application with a custom network protocol over UDP.
Python · Socket programming · UDP protocol · IP routingBuilt SQL data pipelines and automated reporting workflows, improving data accuracy by 10%.
SQLBuilt a PyTorch sentiment classifier for video comments and compared RNN, GRU, and LSTM architectures.
PyTorch · Python · Scikit-learn · Neural networksConverted 7 million song plays from 54 users into matrices for clustering, similarity analysis, and recommendations.
Python · NumPy · Pandas · Scikit-learn · AI algorithms