Senior Machine Learning Systems Engineer - Advertising Infrastructure

Mintegral is a premier programmatic and interactive mobile advertising platform, originating from the APAC region and expanding globally. Driven by advanced AI technology, we deliver innovative, comprehensive solutions to advertisers and developers, helping them achieve their marketing goals through exceptional mobile marketing and monetization strategies. In this role, you will develop large-scale real-time ML inference systems that enhance ad ranking, retrieval, and prediction for global traffic. Ads Serving System ArchitectureEngineer complete ads ranking and serving architectures for real-time bidding and recommendation systems.Create flexible and separated inference pipelines across CPU and GPU layers.Optimize latency for high-QPS ad delivery systems.ML Inference OptimizationDevelop and enhance ML deployment pipelines for diverse CPU/GPU environments.Augment model freshness and inference performance.Facilitate swift iterations of ML models in production.Distributed Systems & Embedding InfrastructureDesign extensive embedding storage and retrieval systems.Create adaptive sharding strategies across different hardware.Strengthen load balancing and overall system stability.Ads Ranking Performance EngineeringOptimize QPS, latency, and throughput.Identify and resolve bottlenecks in inference pipelines.Enhance overall ranking performance from end to end.ML Compiler & Runtime SystemsDevelop AOT compilation frameworks for ML models.Transform models into optimized C++/CUDA/ROCm execution.Elevate inference efficiency across various hardware backends.Cross-functional CollaborationCollaborate with engineers and product teams.Productionize ML models effectively.Establish and uphold standards for scalability and reliability. 4+ years in distributed systems or ML infrastructure.Experience with ML serving or recommendation systems.Solid background in distributed systems and performance optimization.Experience with CPU/GPU systems.Proficiency in C++ / Python.Familiarity with ML frameworks.

Create an account to see the full posting, access our search engine, and more.