Machine Learning Engineer2019 - PRESENTCigna
Technologies: Data Modeling, Splunk
- Engineered various improvements to my KNN model so the solution can identify insurance claims submitted to the incorrect benefit type with 99.8% accuracy overall in 3-5 seconds per claim.
- Yielded the company about $30 million in savings every year with the laser-precision of the solution.
- Created CLI scripts for auto restarting APIs, CLI applications for auto fetching logs from distributed databases, monitoring dashboards with Splunk enterprise, and more as needed.
- Created design docs and schematics to submit with the company patent application for this product.
- Began data exploration and modeling for a new NLP use case I pitched to management after the successful launch of ICC.
Machine Learning Engineer2018 - 2019Express Scripts
Technologies: Zeppelin, Hadoop, Java, Scala, Apache Spark
- Curated the finest tools, processes, and best practices for development on the organization’s first ML team.
- Crafted a binary KNN classifier in Scala with Hamming distance from a training set I translated into binary code. This improved processing time so the model could be used with a streaming solution and externalized spring API in real-time.
- Tested various iterations of my model on different features, training data, and test data with Zeppelin notebook and as a jar on an internal Hadoop cluster to ensure optimal accuracy and that my solution was not overfitting.
- Launched our production pilot to evaluate the performance of my new KNN model, the communication framework around it, and the accuracy of results. I implemented incremental improvements with feedback from our business stakeholders.
Associate Software Engineer2017 - 2018Express Scripts
Technologies: REST API, Oracle, Pega, Java
- Created an enterprise API in Java around a legacy product called Oracle Advanced Queue. It facilitates enterprise wide patient eligibility searches, saving about $100,000 dollars every year. It was the biggest value added FY 2017 on the project.
- Designed and built new Pega applications that automate and manage the lifecycle of a specialty prescription. I got certified by Pegasystems as a systems architect.
- Solved a difficult problem that arose in an MVP1 production release that was impeding our entire project because users were unable to process payer information. I received recognition from my organization’s VP afterward.
- Authored many confluence pages to help our engineers, users, and business stakeholders understand our software. I also collaborated with the end-user experience team to improve productivity across the enterprise.