Top Tech Jobs & Startup Jobs in San Francisco Bay Area, CA

Research and build large-scale pre-training data systems for frontier AI models. Responsibilities include sourcing, curating, filtering, and evaluating text, code, and multimodal datasets; developing quality metrics; scaling reproducible data pipelines; mitigating privacy, safety, and licensing risks; analyzing downstream model effects; and publishing research and tools.

Own the reliability, performance, and uptime of large-scale post-training and reinforcement learning systems. Debug distributed failures across accelerators, networking, storage, schedulers, and training frameworks; build monitoring, alerting, recovery, checkpointing, and scheduling tools; improve cluster utilization and fault tolerance; support production model runs through on-call rotations and postmortems; and partner closely with research teams during active training runs.

Own and evolve security infrastructure for AI foundation-model platforms across compute, storage, networking, and data systems. Architect secure Kubernetes and cloud patterns, manage identity and secrets, secure data pipelines, create threat models, and automate policy enforcement and security checks. Partner with research and engineering teams to deliver scalable systems with strong security, reliability, and performance.

Design, build, and operate large GPU supercomputing clusters supporting large-scale AI training and inference. Responsibilities include cluster provisioning, automation, scheduling, orchestration, capacity planning, storage and artifact management, monitoring, reliability improvements, and performance optimization. The engineer will partner with researchers on scaling runs and advise on distributed training and infrastructure trade-offs.

Design and optimize high-performance ML kernels for large-scale language model training. Develop CUDA, CuTe, and Triton implementations for attention, matrix multiplication, gating, and normalization; improve memory efficiency, low-precision arithmetic, distributed compute, infrastructure stability, and hardware utilization. Collaborate with researchers and systems architects, benchmark performance across GPU generations, and contribute reusable kernel libraries, technical documentation, and open-source or research work.

Coach managers and advise leadership on organizational design, succession planning, team dynamics, and talent decisions. Design scalable compensation, career-leveling, feedback, and evaluation systems for a high-growth AI research lab. Support researcher and engineer performance, development, retention, and growth while ensuring employment practices align with North American law and company values.

Owns the engineering and operation of a large macOS endpoint environment. Responsibilities include MDM platform management and migration, configuration as code, zero-touch provisioning, patching, software distribution, fleet telemetry, compliance, endpoint security, device trust, automated remediation, and advanced troubleshooting. The role also leads technical initiatives, architecture decisions, progressive delivery, and integrations across IT, Security, Identity, and Infrastructure.

Own data projects supporting AI researchers, from understanding requirements and sourcing vendors to reviewing quality and delivering data. Build vendor partnerships, investigate new data domains, create scalable processes and tooling, automate repetitive work, and coordinate with legal and technical teams. The role requires strong judgment, communication, initiative, and comfort managing ambiguity and changing research priorities.

Owns the company’s GRC certification roadmap, including SOC 2, ISO 27001, FedRAMP, and related audits from scoping through closure. Manages control testing, evidence collection, risk registers, policy attestations, regulatory requirements, and audit responses. Partners with engineering, security, legal, safety, and product teams to translate technical and regulatory requirements into actionable controls. Builds compliance tooling, documentation, and a multi-quarter roadmap to scale the GRC function.

Owns end-to-end reliability for Tinker, including CI/CD, observability, service-level objectives, incident response, multi-tenant isolation, resource scheduling, and vulnerability remediation. The role designs monitoring for distributed training systems, improves recovery and checkpointing, and operates Kubernetes clusters supporting heterogeneous GPU workloads. It collaborates closely with platform, security, engineering, and research teams to improve production resilience and utilization.