ML Infra Tech Lead: Scalable Training & Inference

Reducto; San Francisco (CA); On-site; USD 180,000 - 260,000; Full time

Benefits offered by this job

Unlimited PTO

Daily Lunch

Commuter Reimbursement

Comprehensive Insurance

Health and Wellness Budget

Parental Leave

Job summary

Reducto in San Francisco is seeking an ML Infrastructure Tech Lead to own and scale high-performance model training and inference systems. This hands-on role emphasizes building, debugging, and optimizing the production stack while guiding architectural decisions.

You'll work across kernels, runtimes, batching, and distributed training with Kubernetes, collaborating closely with ML and Platform teams. This in-person role requires delivering impact quickly in a fast-moving startup.

Qualifications

• 5+ years building production infrastructure with ML systems.
• Led complex technical projects from ambiguous problems through production.
• Able to set direction and implement the hardest parts.
• Strong Python and systems-engineering skills.
• Understanding of GPU training/inference performance.
• Comfortable with Kubernetes and distributed frameworks.
• Ability to reason across low-level performance and high-level architecture.
• High bar for quality, precision, and reliability.
• Thrives in fast-changing, high-growth environments.
• Takes ownership from strategy to execution.

Responsibilities

• Own the technical direction and roadmap for Reducto's ML infrastructure.
• Build and maintain training and inference stack balancing speed and production.
• Optimize model serving at kernels, runtimes, batching, scheduling, and distribution.
• Design systems for multi-node, multi-GPU training and inference.
• Improve GPU utilization, latency, throughput, reliability, observability, cost.
• Develop benchmarks to identify bottlenecks and guide investments.
• Evaluate advances in training and inference and apply the ones that matter.
• Build tooling and abstractions for moving from experiments to production.
• Partner with ML and Platform teams on architecture and prioritization.
• Raise the engineering bar through design reviews, mentorship, and leadership.

Skills

Python

Systems engineering

GPU training

Kubernetes

Distributed training

Production infra

Leadership

Ownership

Tools

CUDA

Triton

PyTorch

TensorRT-LLM

vLLM

Ray

Job description

Reducto in San Francisco is seeking an ML Infrastructure Tech Lead to own and scale high-performance model training and inference systems. This hands-on role emphasizes building, debugging, and optimizing the production stack while guiding architectural decisions.

You'll work across kernels, runtimes, batching, and distributed training with Kubernetes, collaborating closely with ML and Platform teams. This in-person role requires delivering impact quickly in a fast-moving startup.