Job Description

Anthropic is seeking an Engineering Manager to lead their Inference & Performance Scaling teams. This group sits at the core of Anthropic's operations, focusing on optimizing compute efficiency, scaling large language model (LLM) serving systems, and eliminating infrastructural bottlenecks across inference and training pipelines.

As a technical front-line manager, you will guide direct reports, prioritize high-stakes engineering projects, and ensure that massive compute systems run reliably without compromising Anthropic’s core AI safety commitments.

Platform Labor Intelligence & Technical Stack Analysis

Role & Compensation Context

• Market Positioning: Base compensation of $425k–$560k USD places this role in the top 5th percentile for engineering leadership in the AI infrastructure sector, reflecting high demand for inference optimization experts.

• Technical Complexity: High-concurrency, low-latency distributed systems operating at massive GPU cluster scale.

• Career Trajectory: Engineering Managers in performance and inference scaling typically advance into Director of Infrastructure or Head of Compute Operations roles within 2–4 years.

Technical Environment Overview

• Hardware & Acceleration: GPU/Accelerator cluster optimization (CUDA, Triton, custom silicon execution).

• ML Infrastructure: Deep learning runtime optimization, memory management (KV caching, batching techniques), ML framework internals (PyTorch/JAX).

• Systems Level: OS internals, distributed systems communication, network bandwidth, and memory bandwidth optimization.

Core Responsibilities

• Technical Leadership: Guide engineering efforts aimed at scaling model throughput and reducing end-to-end latency for Claude and internal research workflows.

• Hands-on Contribution: Maintain sufficient technical depth in the team’s codebase and architecture to review designs, solve complex problems, and make targeted contributions as an individual contributor.

• Project & Team Execution: Direct day-to-day work, manage sprints, and set priorities within a rapidly changing, dynamic research environment.

• People Management: Mentor engineers, facilitate performance reviews, design career growth plans, and scale the team through active hiring.

• Alignment with AI Safety: Integrate AI safety principles directly into systems design and infrastructure planning.

Key Qualifications & Requirements

Preferred Technical Background

• Experience with large-scale ML training and inference serving engines (e.g., vLLM, TensorRT-LLM, Megatron).

• Background in GPU programming (CUDA), kernel optimization, or OS internals.

• Familiarity with Transformer architectures and model parallel execution strategies (Tensor, Pipeline, Expert parallelism).

Candidate Guidance & Interview Preparation Tips

• Demonstrate Technical Depth: Anthropic’s leadership style emphasizes technical manager credibility. Prepare to discuss architectural trade-offs, GPU memory bottlenecks, and latency vs. throughput optimization strategies.

• Focus on AI Safety & Mission: Anthropic operates as a Public Benefit Corporation (PBC). Candidates should articulate a clear understanding of why infrastructure reliability and safety go hand-in-hand.

• Emphasize Execution under Uncertainty: Highlight instances where you led teams through rapid growth, managed shifting research priorities, or brought structure to ambiguous systems.

Working Conditions & Benefits

• Office Expectations: Location-based hybrid setup requiring at least 25% in-office time at one of Anthropic's hubs (San Francisco, New York, or Seattle).

• Visa Sponsorship: Visa sponsorship is available for qualifying candidates based on role requirements and regional legal constraints.