Direct reports: 2- Integrated ChatGPT with our virtual benefits assistant, Sofia, in a custom solution to provide LLM-generated responses for more relevant and conversational chatbot interactions. For interactions using the LLM-generated responses, we observed improvements in resolution rates more than 2x the improvements observed for interactions without the GPT integration.- 2024 Hackathon winner: Over the course of a few days, our team built a proof-of-concept of a LLM-based version of Sofia, our virtual benefits assistant, specifically designed to assist our admin users in with common questions related to system configuration, with the potential to save the company hundreds of thousands of dollars in operational savings.

Direct reports: 1- Revamped an outdated version of our site search to integrate Azure AI Search and ChatGPT for a vastly improved search experience, including LLM-generated summaries of search results, leading to a 21.7% YoY increase in search traffic and a 91.6% seven-day resolution rate.- As part of a cross-team effort, I collaborated with the Member Services team to integrate Sofia, our virtual benefits assistant, into the IVR phone menu, more than doubling the number of calls that were able to be fully managed through self-service, substantially reducing wait times and the number of live agent support hours needed.

- In my first 6 months, I worked with the verification team to build a dependent audit product to save a client an estimated ~$1.6M in annual premiums and laid the foundation for additional revenue through the sales of future audits.

• Built automated systems to produce comprehensive monthly profit metrics reports for senior management• Served as the on-team expert for one-off tool creation and data queries• Completed monthly tracking and renewal rate setting for 5 major blocks of in-force business

• Created 2 Python tools now required as part of the routine actuarial model review process• Learned Apache Pig and converted an existing reserve rollforward process from SAS within a single quarter• Completed numerous model changes in MG-ALFA actuarial modeling software• Developed a prototype process in Python for building MG-ALFA input files based on multiple large third-party datasets within a 2-month timeframe

• Coordinated with third-party vendors to implement inbound and outbound single sign-on configurations• Troubleshot dozens of tickets per month submitted by internal customers for access-related questions• Assisted in debugging, maintaining and continuously improving Java web security applications

• Communicated data exchange specifications with assigned HR clients and insurance carriers• Accurately configured hundreds of inbound and outbound data files for the SaaS benefits administration platform primarily using JavaScript and XML• Consulted with internal system architects on system design for dozens of client implementations

• Served as the technical consult on weekly client status meetings for assigned book of 2-5 clients• Ensured the successful exchange of election and demographic data with all third-party benefits providers

• Supported the claims platform by debugging and enhancing existing codebase using the mostly proprietary programming language, MAPPER by Unisys