Technical Program Manager, Inference Performance

Anthropic•San Francisco, WA

7h•Hybrid

About The Position

As a Technical Program Manager for Inference, you'll be the critical bridge between our inference systems and the broader organization. You'll drive strategic initiatives across inference runtime and accelerator performance—coordinating model launches, managing cross-platform dependencies, and ensuring reliability across multiple hardware targets. This role is essential for keeping our most contended infrastructure teams shipping effectively while Research, Product, and Safety all depend on their output.

Requirements

Have several years of experience in technical program management, with proven success delivering complex infrastructure programs, preferably in ML/AI systems or large-scale distributed systems
Have deep technical understanding of inference systems, compilers, or hardware accelerators to engage substantively with engineers and identify technical risks.
Excel at creating structure and processes in ambiguous environments, bringing clarity to complex cross-team initiatives
Have strong stakeholder management skills and can build trust with both technical and non-technical partners
Are comfortable navigating competing priorities and using data to drive technical decisions
Have experience with infrastructure scaling initiatives, hardware integrations, or deployment governance
Thrive in fast-paced environments and can balance strategic planning with tactical execution
Are passionate about AI infrastructure and understand the unique challenges of deploying and scaling large language models

Responsibilities

Systems Integration & Coordination : Lead cross-functional initiatives for new infrastructure integration, establishing clear ownership, timelines, and communication channels between teams.
Drive end-to-end planning for major infrastructure transitions including platform modernization and new tech adoption.
Performance & Efficiency: Partner with engineering teams to identify optimization opportunities, track performance metrics, and prioritize work that unlocks capacity gains.
Coordinate across runtime and accelerator layers to ensure efficiency wins ship without compromising reliability.
Launch Coordination: Drive end-to-end readiness for model and feature launches across multiple hardware platforms.
Establish processes for cross-platform validation, manage launch timelines, and ensure smooth handoffs between runtime, accelerator, and downstream teams.
Strategic Planning: Own and prioritize the inference deployment roadmap, working closely with engineering leadership to prioritize initiatives and manage dependencies.
Provide visibility into upcoming changes and their organizational impact.
Stakeholder Communication: Build strong relationships across research, engineering, and product teams to understand requirements and constraints.
Translate technical complexities into clear updates for leadership and ensure alignment on priorities and timelines.
Process Improvement: Identify inefficiencies in current workflows and drive systematic improvements.
Establish metrics and dashboards to track infrastructure health, capacity utilization, and deployment success rates.