Top AI & Machine Learning Jobs in San Francisco, CA

Develop and operate production-quality AI framework, runtime, benchmarking, performance, automation, and observability systems. Benchmark, profile, debug, and optimize large language model workloads across GPUs and Microsoft hardware. Lead cross-stack investigations, improve model onboarding, reliability, hardware utilization, and Azure capacity efficiency, while collaborating with researchers, infrastructure teams, hardware partners, and engineers. Principal-level responsibilities include setting technical direction, defining architectures, leading strategic initiatives, and mentoring technical leaders.

Lead research and development of AI models for biomolecular modeling and simulation. Responsibilities include integrating molecular simulation and laboratory data, developing deep learning architectures and generative models, writing research software, coordinating large-scale data generation and training runs, and producing technical papers and presentations. The role collaborates with an interdisciplinary research team focused on molecular sciences, computational chemistry, protein science, biophysics, and molecular dynamics.

Drives customer AI transformation by designing and implementing Copilot Chat, Agents, and Microsoft 365 Copilot solutions. Acts as a trusted technical advisor, leads architectural design, applies prompt engineering and agent-building expertise, supports adoption and change management, identifies growth opportunities, and partners with sales and technical teams to deliver measurable business value and sustained customer success.

Design, implement, and ensure quality of systems-level software, tools, and SDK integration for AI frameworks. Collaborate with hardware, ML, and platform teams; write high-performance code (C/C++, C#, Python); define architectures, coding standards, and schedules; and work on device interfaces, runtime systems, and observability/profiling.

Leads the technical vision, architecture, and execution of AI performance initiatives across frameworks, compilers, runtimes, systems, services, and hardware. Drives benchmarking, automation, observability, model performance, reliability, and Azure capacity efficiency. Aligns cross-functional teams, mentors engineers, resolves complex technical tradeoffs, and remains hands-on with prototyping, implementation, debugging, design reviews, and operational readiness.

Design and develop secure, performant Copilot products and platform services across Windows, Mac, and mobile. Build distributed systems and agentic AI experiences that reason, plan, and execute tasks. Collaborate with infrastructure, platform, application engineering, and research teams; ship tested, maintainable code; and improve development practices through AI coding agents, prompt design, and human-in-the-loop workflows.

Designs and validates secure, scalable cloud architectures and AI agent solutions for Microsoft partners. Leads architecture sessions, develops proofs of concept and pilots, provides technical implementation guidance, supports legacy-to-cloud transitions, and helps partners build AI practices, intellectual property, accelerators, and managed services. Collaborates with partner executives, architects, developers, and Microsoft engineering to connect AI models with enterprise data, applications, security, observability, and governance.

Conduct hands-on post-training research for open-weight language models used in cybersecurity. Develop datasets, training objectives, supervised, preference-based, and reinforcement-learning methods; build reproducible pipelines; evaluate model performance, generalization, and security; investigate recursive self-improvement; and deliver trained model artifacts and workflows for red-team partners.

Develop and optimize GPU kernels, compilers, runtimes, and distributed inference systems for high-performance LLM serving. Profile workloads, improve latency, throughput, reliability, and hardware efficiency, and integrate optimizations into production systems. Principal-level engineers lead cross-stack performance initiatives and create reusable capabilities across models and hardware. The role also involves collaboration across model, infrastructure, research, and hardware teams, technical mentoring, and raising engineering standards.

Work on end-to-end performance of LLMs: benchmark and optimize inference on Nvidia/AMD GPUs and Microsoft silicon, build tooling for profiling and deployment, accelerate model porting, and collaborate with internal and external partners to reduce hardware footprint and improve time-to-market.