ML Ops Engineer — Reproducible Multimodal Foundation Models

Veeda AI; Zürich; Hybrid; CHF 120.000 - 180.000; Vollzeit

Benefits dieser Stelle

Health insurance

Zusammenfassung

Veeda AI builds multimodal foundation models for Physical AI and teams across AI, robotics, and embodied intelligence. We are seeking engineers to own tooling for large-scale experiments, ensuring reproducibility, and driving end-to-end pipelines from config to artifact in production-like environments.

Join a small, fast-moving team and help push the boundaries of what’s possible with Physical AI while delivering robust tooling and scalable workflows for the next generation of research and

Qualifikationen

• Bachelor's degree or equivalent hands-on experience in CS/engineering.
• Strong Python and software engineering skills with CI/CD experience.
• Experience operating multi-node training jobs and telemetry-based decisions.
• Experience building reproducible pipelines end-to-end.
• Rigorous evaluation methodology, seeds, sample sizes, and CI validation.

Aufgaben

• Run Lifecycle & Launch tooling: define, launch, resume, and kill runs with config-first tooling.
• Experiment tracking & provenance: bind checkpoints to code, configs, datasets, and container digests.
• Checkpoint registry & lineage: manage retention, format conversion, and artifacts for sim.
• Evaluation in CI: gate checkpoints with seeded rollout and performance suites.
• Goodput & incident response: monitor runs, report GPU-hours usage, and respond to issues.

Kenntnisse

Python

CI/CD

Experiment tracking

Distributed training

Reproducible pipelines

Telemetry

Evaluation methodology

Git

Ausbildung

Bachelor's degree

Tools

GitHub Actions

Buildkite

Argo Workflows

Flyte

Ray

Prometheus

Grafana

OpenTelemetry

Jobbeschreibung

Veeda AI builds multimodal foundation models for Physical AI and teams across AI, robotics, and embodied intelligence. We are seeking engineers to own tooling for large-scale experiments, ensuring reproducibility, and driving end-to-end pipelines from config to artifact in production-like environments.

Join a small, fast-moving team and help push the boundaries of what’s possible with Physical AI while delivering robust tooling and scalable workflows for the next generation of research and