
Rubaiyat
Machine Learning Engineer
Habilidades

Revisa mis servicios

Experiencia laboral
Machine Learning Engineer
Netflix • Tiempo completo
Jun 2024 - Present • 2 yrs 4 mos
Optimized Real-Time Inference Latency: Re-architected the scoring pipeline for the homepage recommendation system, leveraging model quantization and Triton Inference Server to reduce P99 inference latency by 35%, improving UI responsiveness during peak traffic hours. Scalable Feature Engineering Pipelines: Designed and deployed distributed feature pipelines using Spark and Apache Flink, reducing data processing time for offline-to-online feature synchronization by 40% and ensuring near-zero data drift. Cost Reduction in Cloud Infrastructure: Spearheaded a model pruning and distillation initiative for secondary personalization ranking models, cutting GPU cloud compute expenditures by $180,000 annually without degrading recommendation accuracy (Normalized Discounted Cumulative Gain - NDCG). A/B Testing & Personalization Uplift: Collaborated with product and data science teams to design, launch, and analyze 12+ multi-armed bandit and deep neural network (DNN) A/B tests, resulting in a 1.4% lift in user engagement and improved retention metrics across international markets.