Jobs at Vast.ai

Deploys, fine-tunes, and serves open-source AI models on rented GPUs using tools such as vLLM and SGLang. Builds public example repositories, Docker templates, benchmarks, and live endpoints, then teaches through technical writing, videos, talks, livestreams, and community support. Reports product friction to engineering and product teams, represents Vast at hackathons and conferences, and serves as a highly credible technical voice across developer communities.

Design and scale infrastructure powering a global GPU marketplace. Responsibilities include GPU provisioning, workload scheduling, orchestration, provider onboarding, usage tracking, billing, pricing logic, resource management, and marketplace APIs. The engineer will optimize performance, reliability, security, fault tolerance, and multi-tenant cloud systems while collaborating with product and infrastructure teams. The role requires strong Python and C++ development, distributed systems expertise, and experience building large-scale compute platforms.

Lead approximately eight engineers across product and infrastructure teams while remaining hands-on in the codebase. Own platform delivery, engineering quality, reliability, incident response, hiring, team design, performance development, and execution cadence. Partner with the CEO on technical direction and product priorities, ship critical-path software, build a hiring plan for at least four senior engineers, and establish effective release and incident-review processes.

Develop GPU kernels, parallel libraries, tensor libraries, and auto-optimization tools to improve AI inference performance. Research and evaluate state-of-the-art models, frameworks, tools, and system architectures. Contribute to market-based resource management systems while collaborating with technical leadership and engineering teams. The role requires deep GPU architecture knowledge, neural network performance expertise, systems engineering skills, and strong research capabilities.

Develop and extend a GPU cloud daemon that orchestrates hosts across a large fleet. Improve infrastructure performance, reliability, monitoring, telemetry, and containerization. Design market-based resource management systems, harden code and infrastructure to zero-trust standards, and benchmark and eliminate bottlenecks across hypervisor, container, and network layers.

Design and optimize GPU kernels and tensor libraries for scalable AI inference. Apply HPC and parallel-computing techniques, evaluate emerging GPU architectures and resource-management approaches, and improve GPU infrastructure efficiency. The role requires advanced C++ development, parallel programming expertise, systems optimization, and performance tooling experience. This is a full-time, on-site position in San Francisco or Los Angeles.