Bengaluru, Karnataka, India
-- Technical Expertise & Impact: * Refactored Ingestion Layer with Samza, Drools, Resilience4j, enabling scalable, reusable design; boosted throughput and reduced errors by ~50%, improving reliability of downstream systems. * Built a Generic Auto-Action Handler for Multi-Level Review, automating SLA breaches and Ads approvals, eliminating ~57.5K weekly manual review tasks and easing reviewer workload. * Integrated GenAI autoclear workflows (~98% precision, +26pp recall), saving ~15K reviews/week (~14% of public reviews) while maintaining quality standards. * Migrated DSL → Drools (15K+ LOC) to simplify onboarding, enhance performance, and standardize rules; deployed dashboards and proactive alerts for anomaly detection during GenAI ramp-ups. * Delivered ~7% review savings, stronger SLA adherence, and boosted spam detection with +3pp precision and +4pp recall. * Reduced backlog creation via policy-driven routing and optimized queuing, ensuring efficient reviewer allocation. -- Leadership & Collaboration: * Mentored engineers on scalable design, infra migrations, and stress testing; enabled peers to drive complex, high-impact initiatives. * Partnered with various teams across LinkedIn on GenAI workflows, MLR jury consolidation, and spam automation, aligning technical work with business and policy goals. * Authored/reviewed key RFCs shaping ingestion refactors, autoclear workflows, and spam detection designs, ensuring scalability and robustness. -- Engineering Craftsmanship: * Tackled bottlenecks by adopting Samza over Kafka to reduce congestion; improved reliability with classifier ID–based spam delay mechanisms. * Built frameworks to benchmark outputs and drive iterative improvements. * Produced post-mortems, technical reviews, and TTLG sessions, strengthening maintainability and promoting engineering excellence.
-- Technical Expertise and Impact * Led migration from crowd sourcing review to spam review, improving interoperability and labeling efficiency, while integrating machine learning to enable ~3,000 weekly auto-clear cases. * Directed Play 2.6 migration, refactoring 45,000 lines across ~900 files with zero major issues through phased testing, automation, and stakeholder alignment. * Built a configuration-driven inline translation system, raising foreign language coverage to 70% and expanding accessibility of spam review datasets. * Automated content-fetching and implemented proactive monitoring for semaphore misses, preventing outages and strengthening platform reliability. * Delivered scalable Onyx Workbench features including media labeling, reassignment flows, and onboarding of new content sources, enhancing usability across teams. -- Leadership * Co-led Play migration with cross-functional partners, mitigating risks, aligning stakeholders, and ensuring timely delivery. * Mentored teammates on scalable design, testing, and on-call practices; hosted knowledge-sharing sessions on iReview and review systems. * Anticipated risks by flagging underestimations early, revising plans, and ensuring deadlines for critical initiatives were consistently met. * Supported team growth through interviews, hiring, and fostering open communication and feedback-driven culture. -- Craftsmanship * Authored comprehensive RFCs and solution docs for migrations and Onyx Workbench features, ensuring clarity and scalability. * Designed systematic processes for monitoring, TMC-driven automation, and phased rollouts, enabling fault-tolerant deployments. * Delivered thoughtful peer reviews and feedback, reinforcing quality standards and collaboration. * Resolved bottlenecks with innovative solutions, improving efficiency and laying groundwork for scalable future improvements.