Senior Research Scientist/Engineer - AI Infrastructure

Team Introduction: The infra4AI Research and Architecture Team is responsible for the foundational hardware and software systems specifically engineered to support the demanding and often experimental workloads of developing new artificial intelligence models and systems. It serves as the bedrock upon which researchers and engineers create, train, test, and iterate on novel AI architectures, from large language models (LLMs) to specialized neural networks. We are seeking a highly skilled and motivated AI Infrastructure Researchers and Engineers to join our dynamic team. In this role, you will be responsible for designing, building, deploying, and maintaining the robust and scalable infrastructure that powers our cutting-edge artificial intelligence (AI) and machine learning (ML) initiatives. You will work closely with our AI/ML researchers, data scientists, and software engineers to create an efficient, high-performance environment for training, inference, and data processing. Your expertise will be critical in enabling the next generation of AI-driven products and services. Responsibilities The ideal candidate should be an expert in at least one of the following fields to define and design the next-gen AI Infrastructure: - Infrastructure Design & Architecture - Lead end-to-end design of scalable, reliable AI infrastructure (AI accelerators, compute clusters, storage, networking) for training and serving large ML workloads. - Define and implement service-oriented, containerized architectures (Kubernetes, VM frameworks, unikernels) optimized for ML performance and security. - Performance Optimization - Profile and optimize every layer of the ML stack—ML Compiler, GPU/TPU scheduling, NCCL/RDMA networking, data preprocessing, and training/inference frameworks. - Develop low-overhead telemetry and benchmarking frameworks to identify and eliminate bottlenecks in distributed training and serving. - Distributed Systems & Scalability - Build and operate large-scale deployment and orchestration systems that auto-scale across multiple data centers (on-premises and cloud). - Champion fault-tolerance, high availability, and cost-efficiency through smart resource management and workload placement. - Data Pipeline & Workflow Engineering - Architect and implement robust ETL and data ingestion pipelines (Spark/Beam/Dask/Flume) tailored for petabyte-scale ML datasets. - Integrate experiment management and workflow orchestration tools (Airflow, Kubeflow, Metaflow) to streamline research-to-production. - Collaboration & Mentorship - Partner with ML researchers to translate prototype requirements into production-grade systems. - Mentor and coach engineers on best practices in performance tuning, systems design, and reliability engineering.