G
Back to jobs

ML Ops Engineer (Training Pipelines / Computer Vision)

Outsyders

Canada Remote Full-time

Posted 2026-05-14

Check Your Fit Before You Apply

  • ✓ Fit score for this exact role
  • ✓ Call blockers — what might disqualify you
  • ✓ Tailored resume bullets & cover letter
  • ✓ Skills gap & why-you narrative
  • ✓ Interview prep questions
Sign In to Analyze

New users get free credits on signup

Job Description

Job Title: ML Ops Engineer (Training Pipelines / Computer Vision)

Location: Remote, Canada

About the position:

We're looking for an ML Pipeline Engineer to build and maintain the training infrastructure behind our computer vision models.

This role is focused on training pipelines, not model deployment. The goal is to make it fast and reliable for ML Engineers to run experiments—without spending time on setup, debugging pipelines, or wrangling data.

You'll work closely with ML Engineers and dataset teams to ensure training workflows are scalable, reproducible, and efficient across large datasets and multiple experiments.

You'll play a key role in improving how machine learning is built and scaled in production. By reducing friction in training workflows, you directly impact model quality, iteration speed, and overall team efficiency.

Key Responsibilities:

Training Pipeline Development

Design and maintain end-to-end training workflows that support:

data ingestion → preprocessing → augmentation → training → validation → evaluation → model packaging

Responsibilities include:

•Building reusable training workflows for multiple computer vision tasks
•Automating dataset preparation pipelines
•Managing training configurations across experiments
•Supporting distributed training workflows
•Reducing manual overhead for researchers during experimentation

Training Pipelines

•Build and maintain end-to-end training pipelines (data → preprocessing → training → evaluation)
•Standardize workflows so ML Engineers can run experiments with minimal setup
•Improve pipeline structure, reliability, and maintainability over time

Data & Experimentation

•Support dataset preparation, validation, and preprocessing workflows
•Ensure training runs are reproducible and properly versioned
•Implement experiment tracking to support fast comparison and iteration

Performance & Scale

•Optimize training performance (GPU usage, data loading, I/O)
•Support scaling training workflows across larger datasets and compute setups

Reliability & Debugging

•Identify and fix pipeline bottlenecks and failures
•Add validation and safeguards to catch data or training issues early
•Improve logging and observability of training workflows

Collaboration

•Work closely with ML Engineers to reduce friction in experimentation
•Align training pipelines with evolving model and dataset needs

Required Skills & Qualifications:

•4–8+ years of experience in software engineering, ML engineering, or ML infrastructure roles
•Strong Python and software engineering fundamentals
•Experience building training pipelines (not just deploying models)
•Experience with deep learning frameworks (e.g., PyTorch, TensorFlow)
•Experience working with cloud environments (e.g., GCP, AWS, or Azure)
•Experience with Kubernetes (deploying, scaling, and managing workloads)
•Familiarity with experiment tracking tools (MLflow, Weights & Biases, etc.)
•Strong understanding of reproducibility and configuration management
•Experience working with large datasets and preprocessing workflows
•Experience optimizing GPU-based training

Nice to Have:

•Computer vision experience (depth, segmentation, inpainting)
•Experience with image or video data pipelines
•Exposure to VFX, rendering, or production pipelines
•Experience with Ray and/or Anyscale for distributed training and pipeline orchestration

What Success Looks Like

•ML Engineers can run experiments quickly with minimal setup
•Training workflows are standardized and reliable
•Experiments are reproducible and easy to compare
•Data and pipeline issues are caught early
•Iteration speed improves across the team

About Us:

Outsyders are redefining the art of theatrical 3D conversion. Founded by pioneers from three of the most respected stereoscopic conversion studios, our team fuses decades of experience with cutting-edge machine learning to push the boundaries of cinematic immersion.

Our innovations are trusted by the biggest names in Hollywood. We don't just convert films; we elevate them, delivering a premium 3D experience that enhances storytelling, deepens engagement, and sets a new industry standard.