This role focuses on building and maintaining large-scale infrastructure for multimodal generative models, covering the full development cycle from data processing to training, inference, and serving. It involves working with research scientists and product engineers to optimize performance and drive architectural changes for consumer AI products like Copilot.

This role focuses on the co-design and productionization of next-generation AI systems at datacenter scale, optimizing end-to-end performance and efficiency. It operates at the intersection of models, systems software, networking, storage, and AI hardware, influencing accelerator design, system architectures, and large-scale AI platforms. The role involves analyzing real workloads, developing performance models, and partnering with various teams to drive high-impact ideas into production systems. It also contributes to research and the broader community through publications and open-sourcing.

Senior Researcher in AI Systems at Microsoft Research, focusing on AI infrastructure, ML systems, and HPC systems to improve efficiency and scalability across the full AI model lifecycle, including pre-training, fine-tuning, post-training, and inference.

Research Intern role focused on optimizing the performance of Large Language Models (LLMs), involving architecture and inference performance. Requires PhD student status in a STEM field and experience with LLM architecture or inference performance optimization. Preferred qualifications include experience with GPU kernel performance bottlenecks and optimizing compiler architecture.

Software Engineer 2 on the Microsoft Azure AI Inference platform team, responsible for the hosting, optimization, and scaling of the inference stack for Azure AI Foundary models, including those from OpenAI and other OSS providers. The role focuses on designing and implementing core inference infrastructure, improving performance and efficiency for LLMs and GenAI models, and scaling the platform to meet growing demand.

Principal Software Engineer focused on optimizing the performance of AI model inference, particularly LLMs, across various hardware platforms (GPUs, Microsoft silicon). The role involves deep technical work on the AI software stack, from fundamental abstractions to system-level optimizations, aiming to improve efficiency and reduce costs for large-scale AI deployments, including those for Azure OpenAI service.

The AI Frameworks team at Microsoft develops software for training and deploying advanced AI models, collaborating with hardware teams and partners on supercomputers and AI accelerators. This role involves developing and evaluating core algorithmic and hardware technologies for large-scale AI model training and inference, working closely with ML researchers and developers, and with OpenAI on Azure OpenAI service models. The position requires hands-on software design and development skills in languages like Python, C/C++, and CUDA, focusing on LLM optimization technologies, model scripting, and kernel languages.

Principal Software Engineer on the AI Frameworks team at Microsoft, responsible for developing and evaluating core algorithmic and hardware technologies for large-scale AI model training and inference on novel hardware. Collaborates with ML researchers, system engineers, and partners to optimize and scale AI models, build validation tools, and perform software development in languages like Python, C/C++, and CUDA.

Senior Researcher focused on advancing efficiency across the AI stack for generative AI serving systems, spanning models, ML frameworks, cloud infrastructure, and hardware. The role involves algorithmic and systems optimization for latency, throughput, and cost, with a strong emphasis on driving research ideas through prototyping, validation, and production deployment.

Principal Software Engineer focused on optimizing GPU inference for large-scale deep learning models (LLMs/SLMs) within Microsoft's AI-native monetization platform, serving ads, shopping, and Copilot.

Product Manager for Microsoft Foundry's AI Models & Training organization, focusing on platforms and infrastructure for training, evaluating, launching, and operating AI models at scale. The role involves defining product strategy and roadmaps for model training, inference, experimentation, and platform enablement, driving initiatives across the AI model lifecycle, and enabling internal teams and customers to access and adopt models. Requires strong technical fluency, ability to lead cross-functional initiatives, and experience with AI/ML platforms and infrastructure.

Principal Engineer on the AI Core Infrastructure team, responsible for large-scale GPU management infrastructure and inference/training platforms powering Microsoft's AI workloads. The role involves setting roadmaps, designing backend services, and providing insights for customers to monitor, troubleshoot, and scale AI training workloads on supercomputers. Focus on ML infrastructure, distributed systems, and observability.

Principal Software Engineer on the AI Frameworks team at Microsoft, focusing on developing and optimizing software for running AI models across diverse hardware platforms. This includes working on ONNX, ONNX Runtime for high-performance inferencing and training acceleration, and Foundry Local for on-device inference.

Principal Software Engineer role focused on building and supporting large-scale GPU management infrastructure and inference/training platforms for AI workloads at Microsoft. The role involves architecting, designing, and developing core AI infrastructure services and compute, storage, and networking subsystems for LLM training, customization, and inference.

Senior Software Engineer focused on building and optimizing foundational inference engines and APIs for large-scale AI inference across Azure. The role involves improving latency, throughput, availability, and cost for LLMs, working with OpenAI and open-source models, and developing experimentation capabilities for safe and rapid iteration.

Principal Software Engineer focused on building and optimizing foundational inference engines and APIs for large-scale AI inference across Azure. The role involves driving production-grade serving improvements for OpenAI and open-source LLMs, focusing on latency, throughput, availability, and cost efficiency. Responsibilities include making hands-on engine changes, building experimentation capabilities, and designing inference serving architectures to support multitenant AI systems at global scale.

Principal Software Engineer role focused on building and scaling the AI data-plane for LLM inferencing across Microsoft and Azure. The role involves designing, coding, and shipping core serving systems, smart routing, and request distribution for a wide range of LLMs, aiming for reliability, efficiency, and ultra-low latency.

This role focuses on building and optimizing high-performance runtime systems for large-scale LLM inferencing, specifically for OpenAI chat and multimodal AI models. The engineer will be responsible for systems-level optimization, microservice design, and ensuring the latency, throughput, cost, and reliability of AI inference pipelines.

Applied Research role focused on hardware/software codesign for GPU kernel optimizations to improve efficiency of Large Language Models and Generative AI inference. Involves designing, implementing, and optimizing GPU kernels, researching novel optimization techniques, and profiling performance.

Senior AI Software Architect role focused on optimizing AI model performance and enablement on Maia hardware, involving PyTorch, quantization, parallelization, and inference pipelines.

This role focuses on building and optimizing the infrastructure and developer experience for large-scale ML model training and inference, specifically for Microsoft's AI assistant, Copilot. The responsibilities include improving CI/CD pipelines, developing training tools, enhancing cloud infrastructure, and managing model hosting systems for inference and data generation. The role aims to accelerate iteration and improve the quality of AI models powering innovative products.

Research Intern role focusing on ML systems, numeric precision, data types, and compute technologies for AI workloads at Azure scale. The role involves investigating model efficiency through low-precision formats, quantization, ML kernel development, and benchmarking. It aims to inform decisions on compute platforms, acceleration strategies, and system-level optimizations for training and inference of large-scale models.

This role focuses on building and maintaining tools and systems for LLM inference, optimizing compute efficiency, and enabling researchers to run models for various tasks. It involves working with inference frameworks, GPU kernel programming, and distributed systems to improve model performance.

Research internship focused on AI systems and architecture, investigating performance modeling, architectural analysis, and emerging inference mechanisms for large-scale AI workloads. The role involves analyzing hardware, software, and model interactions, developing performance models, and prototyping new inference techniques.

Research intern focused on accelerating large language models (LLMs) by optimizing performance on custom architectures, involving computer architecture and parallel programming.

Research intern focused on optimizing AI inference systems, including LLM inference, KV caching, request scheduling, and GPU orchestration, to improve latency, throughput, and cost-efficiency.

Senior Software Engineer role focused on optimizing large language model (LLM) deployment on Microsoft's MAIA AI accelerators and GPUs. The role involves building software across the stack, including PyTorch, inference systems (vLLM, SGLang), and performance-critical runtime/kernel components. Responsibilities include architecting tensor computation primitives, extending PyTorch for custom accelerators, improving inference stacks, and optimizing kernels for LLM inference and training workloads.

This role focuses on building and scaling Azure's next-generation Model Router, which involves optimizing and deploying LLMs at a global scale. The responsibilities include applying advanced concepts to product needs, working with large-scale computing frameworks for model improvement, deploying and evaluating models in production, and monitoring their performance over time. The role also involves collaboration with product groups, mentorship, and documentation.

Senior Software Engineer on the AI Frameworks team at Microsoft, focusing on developing AI software that enables running AI models across various devices and hardware. The role involves hands-on software design and development in C/C++ for large-scale model training and inference on novel AI hardware, requiring a strong engineering background and the ability to solve complex technical problems.

The Applied Scientist II will design, develop, and ship AI models into the Teams media stack, focusing on real-time conversation products like Teams. This role involves building end-to-end ML systems, from data cleaning and feature engineering to model training, evaluation, and deployment, with an emphasis on optimizing for performance and memory, and updating deployed models based on A/B testing.

The Principal Software Engineer - Performance Tooling role focuses on optimizing the performance of AI models, particularly LLMs, across various hardware platforms (GPUs, CPUs) and software layers. This involves benchmarking, debugging, profiling, and optimizing for large-scale training and inference to reduce deployment time and hardware footprint, contributing to the efficiency of AI services like Azure OpenAI.

Senior Software Engineer focused on optimizing the inference performance of large language models (LLMs) like those from OpenAI, running on various hardware including GPUs and custom Microsoft silicon. The role involves benchmarking, debugging, and optimizing performance to enable efficient deployment at scale for major Microsoft products and Azure services.

Principal Software Engineer for Microsoft AI's Copilot Discover team, focusing on backend platform for agentic vertical content generation, AI model serving, data ingestion, caching, and serving. Powers content across Microsoft products for over 1B customers. Requires strong distributed systems, cloud storage, and data processing experience.

This role focuses on optimizing the performance of AI inference workloads by exploring micro-architectural innovations and validating end-to-end performance. The Principal Silicon Performance Architect will own performance modeling, analysis, and simulation infrastructure, working closely with chip, system, and software architects to drive data-backed design decisions for improved throughput, latency, and efficiency.

Senior Software Engineer focused on performance tooling for AI frameworks, enabling large-scale training and inference of LLMs on various hardware. The role involves benchmarking, debugging, profiling, and optimizing performance for models like OpenAI's LLMs, aiming to reduce deployment time and hardware footprint.

Senior Software Engineer to work on Azure Translator and Language AI services, focusing on building and scaling large-scale AI systems for natural language processing.

The Azure AI Inferencing team is seeking a Principal Software Engineer to lead the architecture and design of a large-scale, high-throughput, low-latency model-serving platform for Azure OpenAI generative models, supporting billions of requests daily. The role involves end-to-end ownership of solution quality, cross-team collaboration, incident response, and championing security, privacy, and Responsible AI.

Principal Software Engineer to build and scale the core serving systems, request routing, and distribution for all LLMs across Microsoft and Azure customers. The role focuses on delivering inference capabilities reliably, efficiently, and with ultra-low latency for a wide range of AI-powered product experiences.

Senior Product Manager for Microsoft's Foundry Inference & Training team, focusing on product strategy and execution for AI model platforms. The role involves owning product strategy for AI model training, inference, experimentation, and platform enablement, evolving model offerings, driving developer-facing experiences, and defining efficiency metrics. Collaboration with engineering, data science, finance, and go-to-market teams is key, with a focus on solutions for regulated environments.

Principal Software Engineer to design and build a Postgres-based database for modern, AI-native, agent-driven workloads within Microsoft Fabric. The role involves innovating on query planning, execution, and storage layers to support high-performance data access for next-generation applications, leveraging open storage formats and engines.

This role is for a Platform Engineering Manager at Microsoft, focusing on building and scaling the AI platform services that power Copilot, Microsoft's personal AI assistant. The role involves managing a team to develop APIs for finetuning, deployment, and core Copilot experiences, collaborating with AI researchers and product teams to bring AI products to life. The focus is on high-performance, secure, and scalable backend systems for consumer-facing AI products.

Principal Software Engineer role focused on building high-performance software for AI capabilities across Windows & Devices. The role involves architecting and building code for deploying ML models at scale, optimizing edge execution, and guiding system-level decisions for inference, memory, power, and security. It requires defining ML infrastructure strategy and has preferred experience in architecting ML inference pipelines for LLMs, local model integrations, and hardware-aware optimizations.

Full Stack Engineer on the MAI Superintelligence Team focused on ML Efficiency & Observability, building capacity management portals and visibility into model performance for ML researchers and executives. The role involves designing and developing features for user interfaces, integrating with backend APIs for training frameworks, and contributing to internal tooling and infrastructure.

This role focuses on the network architecture for AI accelerator platforms, specifically for high bandwidth and low latency networks critical for AI GPU clusters. The Principal AI Network Architect will evaluate, design, and optimize the network stack from hardware to software kernels, influencing Azure product roadmaps and working with state-of-the-art networking labs. The role requires deep expertise in networking technologies and familiarity with AI model execution pipelines.

Principal Software Engineering Manager to lead a team focused on improving the efficiency, scalability, reliability, and cost of the core infrastructure powering Microsoft 365 Copilot experiences. The role involves acting as a coach, guiding technical design, driving performance optimizations in collaboration with research teams, and ensuring the health and availability of live services.

Principal Software Engineer to advance ad-serving infrastructure, focusing on performance, efficiency, and scalability of next-gen model serving and inference platforms for Ads. Designs and optimizes high-performance serving systems and GPU inference frameworks for deep learning and LLM workloads.

Lead and grow a team building and operating world-class research compute infrastructure, including large-scale GPU clusters and agentic development tools, for Microsoft Research globally.

The Principal Product Manager/Architect will define and guide the technical architecture of Microsoft Foundry, an AI inferencing platform focused on reliability, scalability, and efficiency for large-scale GPU fleets. The role involves setting product direction for reliability, GPU fleet efficiency, capacity management, and engaging with strategic customers. Success metrics include platform reliability, GPU utilization, and customer outcomes.