• Part of two member team that trained a BERT model and created service that could take any "GLG Talk" natural language and turn it into a semantic vector, stored in the vector store Qdrant, that could then be used to find similar documents. Allowed the business to drive millions of dollars in revenue by finding relationships and possible points of revenue by finding similar tasks and opportunities across the business. All the while improving accuracy of engagements. Was deployed as the backing algorithm behind a rerank and automated recommendation system. Also used as a RAG solution for generative AI services.
• Built a mission critical event based service that acts as the fully automated source of truth for which employee is credited with and receives commission for recruiting.
• Refactored and rebuilt the existing search API to have a layer of abstraction between it and the backing SOLR index. This allowed for request validtion, typing via Typescript, and to move away from a single backing data source.
• Maintained and updated search micro services built on top of ElasticSearch
• Helped ETL hundered of millions of rows of data into needed inputs and outputs to train ML classifiers and regressors for revenue forcasting.

• First quarter while employed won the quarterly performance award for the development department.
• Worked on a wide array of applications. From a Ruby on Rails application to a Java Spring application to a ES6 React application and a Typescript React application.
• Helped build a reactive microservices that listened to a number of Kafka events to act as a data pipeline to stitch together customer statements.

• Headed up implementation of a new property search API built on top of ElasticSearch. Worked closely with stakeholders to create an API that met all parties needs.
• Made sure the API was composed in such a way that new team members could be easily onboarded to write code and specs. For instance time for writing a new property filter was reduced from days of a senior developer's time, to an hour of a junior developer's time.

• Took leadership of Upstream's donation platform.
• The platform received an `A` grade from The Washington Post.
• In the 2016 election the platform was one of the 6 candidates, out of 23, that got honor roll for privacy and security.
• Took app to full PCI compliance both code and paper work. Along with increasing test coverage to 90%.
• Built A/B testing system, and built out custom analytics for their donation platform.
• Created an embeddable iframe form system that had a life cycle callback system.
• Helped maintain numerous Ruby on Rails apps with baking stores from MySQL, Postgresql, ElasticSearch.