Software Development Engineer II, Amazon

August 2020 - Present

 

At Amazon I work on the Denied Party Investigation (DPS) team, which screens all customer address for Amazon Retail and AWS to ensure the company is not transacting with a denied party. I specifically work on the investigation platform and decision automation. Meaningful projects I have lead include: 

• A migration of API clients from a 3rd party Investigation Platform onto the DPS Investigation platform to meet the 2020 CFO Goal

• Design and Development of a Machine Learning Service for real-time suppression of DPS Investigations to achieve 2021 Org Goal of 30% investigation automation

• Architecture and Anvil Certification for an Upstream Alert Parsing Service to de-couple an existing monolithic tool 

In my time at Amazon, I have been recognized by Tony Masone (Amazon Treasurer) and John Keimeg (Director of DPS) for my work. I also:  Lead Operational Excellence Syncs for DPS Engineering (40+ engineers), Mentor 4 new hires, and Actively Participate in Interviews.

​

Machine Learning Engineer, IBM

July 2018 - August 2020

 

​At IBM I work on a team that handles all inbound communication on product web pages. I specifically work on deploying, analyzing, and scaling chatbots to enhance the customer experience. Meaningful projects I have worked on:

• Memory and Computationally Efficient Method for Scaling Intent and Entity Data for the Deployment of Foreign Language Chatbots

• Routing Logic for Non-Domain Specific Webpages (eg. IBM.com)

• Question Suggestion Engine to Advance the Customer Journey within Chatbot Conversations

• Chat Transcript Analyzer - Tool that Gains Meaningful Insight from Chatbot Conversations

• Tool to Suggest Companies for Mergers & Acquisitions based off feature extraction from news articles and sentiment analysis

• Mission Statement for the Cognitive Applications Business Unit

At work, I am also passionate about: The Digital Growth and Commerce Engineering Guild | Patenting (5 patents in the NLP space) | Speaking at conferences (Grace Hopper 2019, Women in Statistics and Data Science 2019, International Workshop on Machine Learning and Artificial Intelligence 2019) | Hosting workshops for non-profits (eg. Girl's Who Code, P-TECH)​

​

Core Development Intern, InterSystems

​May 2017 - August 2017

 

At InterSystems, our project focused on showcasing the scope of the company's new database architecture. With the Amazon Reviews Dataset loaded with the InterSystems IRIS Data Platform, we use Natural Language Processing and Machine Learning for feature extraction and feature clustering. After extracting product features from product reviews, we used sentiment analysis to determine the most liked and disliked features of a product. With these features we then generated a concise and meaningful summary of the product. We demoed this project via a user-friendly WebApp, which won Best Team Project at the Intern Showcase. 

​

Teaching Assistant, Carnegie Mellon University

​August 2015 - December 2017

 

For three years I was a TA for EUREKA! Discovery & Its Impact a first-year seminar that sets students up to be successful students and scientists/ mathematicians. As a teaching assistant, I worked with Dean Ken Hovis to lead recitations that equip new students with foundational knowledge, skills and perspectives for supporting their development as scientists. 

 

Research Assistant, Carnegie Mellon University

​August 2014 - May 2018

Identification of Predictive Properties for Malignant, Metastatic Tumor Phenotypes

• Derived feature vectors using Principle Component Analysis to reduce the complexity of inferring cancer evolution

• Used Machine Learning Classifiers and Regression Models to test these vector’s ability to predict tumor progression

• Presented this project as my Senior Honors Thesis

 

Data Imputation for Phylogeny-Based Cell Lineage Reconstruction

• Proposed a new method for imputation based on phylogenetic inference. Validated the method using

• Validated the method by comparing it to genetic alternatives using K-nearest neighbors and Linear Regression

• Contributed to an Open Source Project