Developed and owned backend services for large-scale user-facing web applications

Maintained and enhanced online Ads system with multi-billion dollar revenue

Worked with cross-functional teams to help professionals advertise their business, implemented new web components, and conducted A/B Tests for better ads serving

Elevated the Leads Quality through spam-level analysis utilizing Message Queue(MQ) and the Data Warehouse / Big Data

Identified performance improvement opportunities and cut the runtime by half through Solr and Ads server(Java) adjustments, leading to 5% retention increment

Built and guided several end-to-end internal tools, starting from the system and data (MySQL) design docs, service platform(Thrift & PHP) to test automations; 3+ web teams benefited from these tools

Mentored peers in tasks and career growth

Designed and built a React/Redux web application to monitor trading activities

Worked in a fast-paced, agile environment consisting of a cross-functional team implementing and deploying user email system with internationalization

Created detailed technical designs, code, and unit tests for custom applications and data flows in the context of projects, releases, and production support

Proactively innovated solutions to improve the security of RESTful APIs for KYC approval and MFA, provided technical estimates for user stories, articulated any technical risks and assumptions

Performed code reviews and debugging, ensured code quality and reliability through unit and integration tests

Established productive working relationships and maintained effective communications with teammates, clients, and end-users

Integrated the AR.Drone’s driver and OpenCV through Robot Operating System Framework to automate the image collection

Designed classes and objects for image collection under category of ImageNet and ModelNet

Utilized ResNet, AlexNet and VGG to run experiments on dataset, ImageNet and ModelNet

Participated in design discussions and review sessions with team members, submitted paper to European Conference on Computer Vision in March 2018

Decreased the size of databases using Set Cover and the concept of domain-specific synthesis

Devised and Implemented tags to organize the databases based on unit-selection synthesis

Applied Dynamic Programming to efficiently retrieve the data and achieve high naturalness

Incorporated digital processing to better improve output quality

Identified and suggested improvements in methods, planed project of limited duration and scope, troubleshot when erroneous or unexpected results occurred