Paul Mccabe Email and Phone Number

• Designed, implemented, and supported the companies first Azure OpenAI platform as part of a two person team, with an emphasis on high availability, security, and flexibility. This included provisioning a network of cloud resources, a model-provider agnostic ASP.NET application, and a client for programmatic access.

• Educated dozens of users and made hundreds of configuration changes as a Splunk Admin to support company efforts in a variety of different domains. (Fraud monitoring, conveyor performance, website errors, etc...) This included fixing complex dashboards, locating incomplete data, and resolving performance issues with Splunk Enterprise. • Investigated and resolved off-hours issues with our Azure OpenAI platform such as OpenAI outages, errored requests from our users, and unusual billing activity. • Provide off-hours support for a critical search engine pipeline, ensuring internal teams can perform quality assurance for website updates, directly enabling their day-to-day operations.

• Refactored legacy code across all levels of the search engine stack (frontend JavaScript, backend logic, and Elasticsearch data), proactively identified and implemented improvements beyond project scope, and validated changes through 50k-search comparisons, gradual rollouts with monitoring, and in-depth code analysis.

• Facilitated meetings with Booz Allen SME's to understand and solve an inefficient document conversion process from Visio to MagicDraw format.• Created a Python script and Java plugin to reduce conversion time by up to 70%. • Collaborated with a team of 3 other interns on coding tasks and product development and presented our results to upper management.

• Developed and tested a ML model for patient mortality risk using the LACE index framework. This index can be used by hospital staff without a computer to determine patient risk and improves upon the industry standard's model by 5%.• Developed an ETL pipeline of patient bed alarm data using Python and PostgreSQL. This cleaned data will be used for predictive modeling and reducing false alarms, improving the patient experience.

• Developed an algorithm that uses CT scans to model the airway, bronchioles, and cancerous growths separately in 3D.• These models can be used to predict diagnostics such as malignant vs benign tumors and prevent unnecessary cancer treatments, recognize tumors early, and model the growth rate of tumors.

• Designed wiring diagrams and setup documentation of a coolant monitoring system for the National Astronomical Observatory of Japan, specifically a 50-million-dollar instrument known as the Prime Focus Spectrograph.• Collaborated with telescope faculty to assemble and test 23 sensor-monitor modules to confirm accurate readings.• Installed all modules at the summit, ensuring instrument functionality and measurement accuracy. Data collected will advance our current understanding of Dark Matter and Dark Energy in the universe.