Arun Ajay

Insights & Optimization Team - Playback Org Our team optimizes the Audible playback experience through data-driven insights. Operating across Android, iOS, and Web, we architect and own the ML-focused infrastructure necessary to identify and deliver performance optimizations across the entire playback ecosystem.

Royalties & Accounting Team - Product Design & Engineering Org - Built and optimized features for the royalties and accounting platform, contributing to the system’s scalability and performance. System processes $1B+ in revenue across 55k+ music labels and 15B+ media streams. - Modernized and refactored legacy React applications, improving code quality and driving better usability for internal users. - Designed and implemented backend services and data integrations using GraphQL, Kafka, and Neo4j to enhance data reliability and efficiency. - Led engineering initiatives by defining technical scopes, collaborating cross-functionally, and improving system monitoring and incident response.

Ad Fraud Prevention Team - Monetization Org - Strengthened Twitch’s ad fraud prevention systems by designing and implementing detection mechanisms to safeguard the platform’s integrity. Protected over 5M+ live ad impressions and preserved $100k+ in creator payouts. - Orchestrated a data pipeline processing over 100M+ rows of global Twitch traffic daily, streamlining delivery to Amazon’s Traffic Quality team via Airflow. - Migrated 10% of all Airflow DAGs at Twitch from a legacy Apache Airflow cluster to Amazon MWAA, improving reliability and security. - Created monitoring frameworks using CloudWatch, PagerDuty, and Grafana to improve system observability and incident response across the ad fraud team. Actively monitored 100+ KPIs

Consumer Line of Business Team - Automated internal data workflows by developing Python and Bash scripts to streamline data transfers and validation processes. - Supported cross-team projects by gathering technical requirements and ensuring smooth communication between stakeholders and engineering teams.

• Password Reset IVR - Created an IVR that performs several security checks prior to resetting a user’s password on Active Directory using LDAP, Python, DynamoDB, AWS Connect, and AWS Lambda • Cloud Desktop Serverless API - Revamped the cloud desktop provisioner server to a serverless architecture. Expanded the serverless API with additional calls to generate forms and data using AWS API Gateway and Python • Termination Survey - Created customized, time-sensitive monthly surveys alerting users that their underutilized cloud desktops will be terminated unless requested otherwise using React, AWS API Gateway, Python, and DynamoDB • Cloud Desktop Data Collection - Created a script executed daily to collect additional information about a cloud desktop including billable hours, active directory profile, and cost center via Python, AWS Lambda and DynamoDB • Cloud Desktop Metrics Dashboard – Created an API that powers nearly a dozen visuals on Redash that highlights the cost and usage rate of the cloud desktops on a branch, cost center, or user level via AWS API Gateway and Python • Storyline IVR Metrics – Created a script that analyzes all interactions of the storyline IVR and sends out a weekly email with a CSV detailing each audio file’s hit frequency using AWS Lambda and Python

• Sentiment and Tag Analyzer – Scraped over financial and regulatory documents to build an NLP language model (corpus) to clean and process documents into the proper format based on exploratory data analysis. Trained models that generated a weighted sentiment score for tags based on a category using Python, Spacy, Gensim, and TextBlob

• Tenable-Jira Ticketing System - Created a ticketing system that processes scan reports streamed from Tenable REST APIs, reduces the ticket load from several thousand to approximately a dozen per scan type, and converts them into tickets through Jira REST APIs using Python and SQL • Desktop Provisioner Server - Built a server using Flask that processes HTTP requests for the creation, termination, and status check of AWS Workspaces (desktops) and was deployed on AWS Elastic Beanstalk using Python • Enhanced Image Search Demo - Developed a full-stack web application that simulated an enhanced search query of images from NYPL’s Digital Collections using React, Python, SQL. Built a server, using Flask, to supply the front-end with images filtered by the metadata extracted from AWS Rekognition • Branch Hours Retrieval for Chatbot – Created a function within AWS Lambda that was designed to control the dialogue flow between the user and the chatbot, AWS Lex, in order to retrieve the branch hours of a location using Python • SMS/Voice Alert Interface Demo - Developed a full-stack web application that enables users to send customizable messages via SMS/Voice via AWS Pinpoint to multiple numbers using React and Express.js. Authenticated users prior to accessing the platform with Google using the OAuth 2.0 API via the authentication middleware, Passport.js

• Automated Jargon Report Handler - Built a full-stack web application that enables analysts to query reports and automates the extraction, tagging, and delineation of industry jargon using Angular, Python and SQL

Computer Science Achievement Award - Awarded to four graduating seniors with the highest GPA.