AR Dev Infra

Designed and built a generic client-side buffering and request batching utility in the AWS Java SDK that improves batch request usability and reduces expenses by up to 10x for customers across many AWS services. Hand coded an AWS SQS implementation to lift a blocker for users to transition from SDK 1.x to SDK 2.x. Implemented a fully-scalable code-generation model that automatically generates the request batching utility for all AWS services that can leverage batch requests.

Built a rules repository data module from the ground up using Java and Maven to unify datasets and standardize all business rules, arming stakeholders with a consistent view of data and confidence in verified information. Deployed rules repository to data lake, increasing data consumption and distribution speeds from once-a-day to near real-time. This creates more impactful business decisions informed by the latest available information. Greatly reduced need for costly downstream data repairs by creating an automated workflow that handles all data extraction and processing at information source with ability for bulk data imports and exports.

Used React, HTML, JavaScript, and CSS to migrate Columbia’s largest digital news publication onto a new tech stack. Developed responsive components, such as a news carousel, using bootstrap that fit product design specifications.One of three developers creating the summer fellowship program’s summary section from scratch

Worked as a data scientist at a Columbia University startup using python and other data analysis skills.

Worked on a project to build a machine learning application to detect and classify hate symbols from images in real-time to identify these groups during protests. Trained Faster RCNN object detection models on Open Image Dataset relevant symbols in most-likely objects. Modified TensorFlow object detection API tutorial code and wrote python scripts to be used for symbol detections. Worked extensively with python and the Python Imaging Library in developing scripts that created synthetic data to be used to train machine learning models. Gained experience developing code in a team-based environment, learning to use git and Github alongside other developers. Helped develop preliminary web-app with flask and docker for proof of concept of the project goal.

Worked with top technology companies and connected with engineering managers to deliver high performing candidates for niche and high demanding roles.Optimized recruiting pipelines by collaborating with leadership and building extra functionalities for internal tools.Proactively identified a problem in an essential tool and created a solution that increased user responsiveness by 50%.