• ML Infra, Indexing Platform and Search Infra• Scaling ML Infra(offline, training, inference) within Search and Discovery• Working on everything related to Coupang search index: ground truth data ingestion, ranking signal aggregation, ML reranking data serving, Solr batch and dynamic index building and serving, etc• Building/Optimizing Spark jobs to perfection in terms of time and resource usage• Scaling indexing building and serving to accommodate business hypergrowth• Reducing AWS cost by millions

• Tech lead on Jobs Foundation team• Project Abacus for Jobs – Designed, scoped and implemented an entirely new job fulfillment flow which lists millions of paid jobs on LinkedIn daily – Collaborated and influenced Payments team to tailor storage and API design for LinkedIn Jobs – Migrated both versions of LinkedIn Job Posting Partner APIs w/ zero down time• Bing Geo Location Data Migration for Jobs – Led the design, scoping and implementation effort of the project – Built foundational pieces including dual write, forward and backward backfill and an automated periodic data backfill pipeline – Helped migrate job search, cache updater, and standardization pipeline• Project Firefly – Designed, scoped and led the implementation effort of purging obsolete jobs data from online storage while keeping them available for offline analysis – Built an end-to-end scraped job deletion pipeline which purges up to 10 million jobs per day using Apache Kafka, Gobblin and Lumos, an internal system to ETL data from various sources – Helped rebuild the internal standardization pipeline and saved job feature on LinkedIn – Built new offline data views to minimize impacts to data scientists working on various analysis – Reduced Oracle storage by 80% and achieved 3x speedup in index build time, 3x speedup in DB write latency and reduced replication latency alerts by 98% via partitioning• Project Best – Co-led the design, scoping and implementation effort of splitting critical Jobs functionalities from 2 monolithic apps into multiple micro-services – Created 5 deployables and numerous libraries based on ownership and functionality and migrated production traffic with zero down time all within 2 quarters – Enabled CI/CD for all 5 deployables by increasing coverage and adding extensive integration tests – Significantly improved developer productivity by reducing commit-to-publish cycle by50% to 85% and saving 200 eng-hours per quarter

- Created Zenefits email product, implemented email priority handling by introducing different levelsof priority and queues, enabled email status tracking by integrating with SendGrid api’s- Created internal insurance related auditing architecture for all lines of insurance coverage(includingmedical, dental, vision, life and disabilities) to diagnose missing/bad data/status and trigger corresponding internal required actions for operations people- Built and maintained bulk uploader for plans, rates and region mappings for all lines of insurancecoverage and other various crucial data- Built/upgraded and maintained all major rate-related client and internal products from end to end- Drove major upgrade on company administrator insurance plan shopping product- Redesigned and built new insurance enrollment frontend and backend flows- Created insightful data reports from Redshift via Mode Analytics- Improved product flow by analyzing metrics and using A/B testing- Built a end-to-end seamless in-product customer feedback survey framework using SurveyGizmo, fullyautomated survey creation, status tracking and feedback collection to improve NPS

Oracle Text Development Team, competing with MongoDB

Teaching Assistant of Introduction to Computer Engineering

Worked on GNU Radio with USRP/USRP2 boards. Coded in C++ and Python

Teaching Assistant of Introduction for Signal Processing

Participated in the Metropolitan Area Network core equipment installation project,Installed Huawei NE80E,Learned basic working principles of ASP.NET