Slashed client onboarding time 30% by implementing an ELT system to extract advertising data for different channels and transform them in various formats that powers generating reports for customer centric data analysis

Enhanced backend module communication by building a RESTful webservice in Java that refreshes Airbyte data pulls and process channels on schedule, designed and coded API endpoints that allows data syncing off-schedule when needed

Upgraded data processing service which leverages Java Spark SQL to load and transform advertising data from AWS S3 buckets and persists them into snowflake datastore which helps streamline the data processing flow

Minimized human intervention by scheduling jobs in Python using AWS Lambda and AWS batch to invoke data processing

Provides real-time monitoring and insights into zeenk's backend platform by installing snowflake dashboards

Redesigned and architected a NodeJS web application for user invitation and login flow which served as the foundation for user authentication system

Unblocked critical client's onboarding by setting up machine-to-machine authentication which allows automated machine autentications and configured Okta as SAML identity provider to allow user authentication with existing third-party Okta account

Implemented machine-to-machine user authentication flow that allows backend to backend interaction and configured Okta as SAML identity provider to allow user authentication with existing Okta account

Designed and tested REST API to ensured smooth transition for legacy users from existing PHP system into new user authentication system. This migration signicately reduced cost and failure rate

Migrated and re-implemented legacy RTB(Real-time Bidding) Platform into Java Dropwizard-based webservice which expedited code deployment and overall system response time

Worked with team using agile and scrum development methodologies

Wrote a script perl which processes regression data to streamline GRS flow, enabling minimal or no change when adapted to other regression progress

Applied mySQL database to store regression result and XML file as interface to specify display rules and PHP to display regression result on internal website for easy auditing and sigh-off

Created a file parser to parse different domains of PrimeTime timing report files in python which acts as the key part to help bring productivity to the timing debug process

Programmed a TAP viewer which extracts, organizes and integrates both timing and physical data in a single environment and provides multiple ways to navigate through the timing report