Software Engineer II (Borrower Acquisition Team)

· Developed an ML-driven communications engine using Temporal, Aurora Serverless, AWS MemoryDB, and Airbyte, enabling personalized email and SMS outreach for over 2 million accounts.

· Led scaling effort from 30k to 4M+ accounts with a 10x growth target, generating estimated $5M for the company.

· Pioneered communications engine in a greenfield Modular Monolith architecture, collaborating with product and Developer Experience teams to establish local, QA testing, and CI/CD environments—the first initiative of its kind within the engineering organization.

· Collaborated with Product members to design & develop feasible MVP criteria, feature flag rollout, and modeling experimentation framework with the communications engine.

· Performed Disparate Impact Analysis on ML Models used to drive company personalization strategy by analyzing SHAP values and DII indexes. Created a generalizable notebook for Analytics team to evaluate future models with, driving trust and safety efforts amongst clients + borrowers.

Software Engineer I (DevX Team)

· Spearheaded SCIM API integration with Okta as tech lead, allowing clients to self-service user accounts, helping close a multi-million dollar deal with a client, and saving >30h/client on manual user management

· Collaborated on splitting up Frontend dependencies and deployments using NPM/React, AWS CodePipeline/Build, and AWS Lambda, dropping deployment times by 67.5% and allowing FE teams full ownership of their React code and dependencies

· Developed a searchable datastore as an audit service for Zero Touch Publishing team to investigate title publishing misses or takedowns on the Prime Video catalog using AWS Lambda, SQS, and OpenSearch

· Implemented features including statistics, graphics, and information displays of missing data that lead to takedowns on the catalog

· Delivered beyond project-scope business value through statistics, graphs, and metrics features to allow for future data-driven decisions on iterations of existing systems in the Prime Video pipeline

· Optimized investigation resolution time ranging from ~5h-2 days on average to <2 minutes by enabling simple UI-based queries on the data for catalog gating decisions on specific titles using OpenSearch Discover

· Planned, taught, and judged hackathons for 60+ middle-school students in Atlanta to learn how to code and create musical submissions in Python for the Amazon Earsketch Competition

· Aided Amazon Future Engineer in the collaborative effort for a 7 week set of Ciara Song Remix Weekly Challenges for k-12 students over the nation by filming 4 introductory videos and creating prize categories for each challenge

· Designed a phone app on AdobeXD to teach over 4,000 k-12 participants of the k-12 Inventure Prize Competition the basics of idea development

· Established creativity building game for 60+ 4-8 grade students using theories on educational gamification, NLP, and machine learning to analyze student answers to creative prompts

· Built an end-end Self-Service UI tool for internal customers to gather information on changes to titles on Prime Video through web service offering versioned data via React, AWS DynamoDB, Lambda, DynamoDB Streams, Kinesis Streams, API Gateway, S3, and CloudFront Distributions

· Improved Offer Distribution team’s 4-18 day root causing failures rate to <1 day to mitigate offer change issues at an efficient scale

· Implemented a cache warmer for Prime Video using AWS CloudWatch Insights, Lambda, SQS, DynamoDB, and DAX to prepare customers for DAX cluster replacement due to the 4-500k steady state through-put of single global catalog (One-Catalog) initiative without compromising their 40ms data retrieval time limit

· Optimized an initial 100% miss rate on the cache to 100% hit rate, allowing downstream customers to retrieve data in <40ms/request