Junior Computer Vision & ML Infrastructure Engineer

Obruta Space Solutions Corp.; Ottawa; Hybrid; CAD 60,000 - 100,000; Full time

Benefits offered by this job

Health & Dental

3 weeks paid leave

Stock options

Job summary

Obruta Space Solutions Corp. is seeking a Junior Computer Vision & ML Infrastructure Engineer (new graduates welcome) to elevate our 3D vision training, evaluation, and benchmarking pipelines. You will design robust MLOps automation for large-scale training across cloud GPUs and local workstations.

You will help benchmark and verify 3D pose estimation models under space-like conditions, integrating datasets from Unreal Engine/Polar and real lab data while coordinating with flight software teams.

Qualifications

• Bachelor's or Master's degree in Computer Science, Computer Engineering, Electrical Engineering, Robotics, or related field.
• Strong programming in Python and familiarity with PyTorch.
• Hands-on experience building, training, and evaluating CV/ML models.
• Foundational software engineering: Git, testing, Linux environments.
• Knowledge of linear algebra, 3D geometry, coordinate transformations, and camera fundamentals.
• Authorized to work in Canada and able to obtain security clearance under Controlled Goods Program.

Responsibilities

• Build, automate, and maintain scalable training, data preprocessing, and model checkpointing workflows in Python and PyTorch.
• Develop automated evaluation harnesses to benchmark 3D vision and 6-DOF pose estimation models.
• Design regression tests across synthetic and real data for robustness.
• Automate data ingestion, augmentation, and format conversion for benchmarks.
• Assist in design and fine-tuning of 3D vision models and event-camera algorithms.
• Validate model exports (ONNX, TensorRT) for embedded edge targets.
• Create automated test reports and visualizations for verification packages.

Skills

Python

PyTorch

Git

Linux

3D geometry

Camera fundamentals

Education

Bachelor's or Master's degree in Computer Science / Engineering

Tools

OpenCV

Open3D

Docker

Weights & Biases

MLflow

GitHub Actions

Job description

Location: Ottawa, Ontario (Hybrid / Remote-First with periodic work at the Ottawa facility)
Position Type: Full-Time
Compensation: Salary & Stock Options
Benefits: Comprehensive Health & Dental package, starting 3 weeks paid annual leave
Contact: If you have issues with your application, please email [e-mail]

About Obruta

Founded in 2019, Obruta Space Solutions is enabling the in-space economy by developing the advanced autonomy systems required for in-orbit satellite servicing, logistics, and active debris removal. Today’s satellites are single-use assets that cannot be refueled, inspected, or repaired. Obruta is eliminating this barrier with our flagship product: the RPOD Kit —a turnkey, vision-guided Guidance, Navigation, and Control (GNC) subsystem that makes autonomous spacecraft docking commonplace.

Backed by leading venture capital and top-tier space innovation programs, Obruta is executing cutting-edge perception and navigation projects with the Canadian Space Agency (CSA), National Research Council (NRC), and international commercial partners.

The Role

Obruta is seeking a motivated Junior Computer Vision & ML Infrastructure Engineer (new graduates welcome!) to take our computer vision training, evaluation, and benchmarking pipelines to the next level of engineering rigor.

In this role, you will be the engine behind how we train, validate, and verify our 3D vision and 6-Degrees of Freedom (DOF) pose estimation models. Instead of manually running training scripts, you will design and implement robust MLOps automation for large-scale training runs across cloud and local GPU compute. You will build comprehensive automated benchmarking suites to evaluate model accuracy, latency, and failure modes across both synthetic Unreal Engine/Polar datasets and real-world test data from our 6-DOF dual-arm robotics laboratory.

You will also help benchmark and validate algorithms operating under challenging space conditions—including extreme high-contrast lighting, noisy sensor feeds, and novel neuromorphic (event-based) data streams—establishing reproducible performance baselines that give our flight software and GNC teams complete confidence.

Key Responsibilities

• Automated Training Pipelines (MLOps): Build, automate, and maintain scalable training, data preprocessing, and model checkpointing workflows in Python and PyTorch across local workstations and cloud GPU instances.
• Rigorous Benchmarking Suites: Develop automated evaluation harnesses to benchmark 3D vision and 6-DOF pose estimation models against standard metrics (e.g., ADD-S, translation/rotation errors, edge-case failure rates, inference latency).
• Sim-to-Real & Robustness Testing: Design regression testing frameworks that continuously evaluate algorithm performance and domain generalization across synthetic datasets (Unreal Engine/Polar) and empirical optical/neuromorphic imagery collected in our dual-arm robotics lab.
• Dataset Management & Curating: Automate data ingestion, augmentation, format conversion, and quality-screening pipelines for synthetic data streams and benchmark datasets.
• Perception Algorithm Development: Support the design and fine-tuning of 3D computer vision models, classical feature-matching pipelines, and novel event-camera algorithms under the mentorship of senior technical leads.
• Model Export & Deployment Awareness: Validate standard model exports (e.g., ONNX, TensorRT) against ground-truth baselines to ensure precision is maintained, keeping downstream embedded edge targets and compute budgets in mind during model design.
• Verification Reporting: Generate automated, repeatable test reports and performance visualizations to support technical deliverables and commercial verification packages.

Required Qualifications

• Bachelor's or Master's degree in Computer Science, Computer Engineering, Electrical Engineering, Robotics, or a related discipline (recent graduates encouraged to apply!).
• Strong programming proficiency in Python and familiarity with deep learning frameworks (preferably PyTorch ).
• Hands‑on academic, personal, or internship project experience building, training, and evaluating computer vision or machine learning models.
• Solid grasp of software engineering fundamentals: modular code design, Git version control, testing best practices, and Linux environments.
• Foundational understanding of linear algebra, 3D geometry, coordinate transformations, and camera fundamentals.
• Applicants must be legally authorized to work in Canada. Successful candidates must able to obtain and maintain a security assessment under the Controlled Goods Program.

Nice-to-Have Qualifications

• Sim-to-Real & Transfer Learning: Exposure to sim-to-real transfer techniques, domain adaptation, domain randomization, or fine‑tuning pre‑trained representations.
• Challenging Environments & Uncertainty: Familiarity with vision algorithms handling extreme lighting (harsh solar glare, deep shadows, high dynamic range), low‑signal/noisy sensor data, or state estimation under uncertainty.
• MLOps & Workflow Tooling: Familiarity with experiment tracking and automation tools (e.g., Weights & Biases, MLflow, Docker, GitHub Actions).
• Classical Computer Vision: Experience with classical computer vision libraries ( OpenCV , Open3D) or basic C++ programming.
• Synthetic Data Pipelines: Exposure to synthetic data generation tools (Unreal Engine, Blender, Unity).
• Deployment Awareness: Basic familiarity with edge deployment concepts (e.g., ONNX, TensorRT, inference latency, memory footprints) to help design models with embedded edge hardware (like NVIDIA Jetson) in mind.
• Novel Sensing: Exposure to or interest in neuromorphic sensing / event‑based cameras.

Why Join Obruta?

• High Impact Early in Your Career: Build the automated foundation that trains and validates vision models destined to fly on autonomous spacecraft.
• Mentorship & Growth: Work closely with experienced engineers across GNC, embedded flight software, and space robotics while maintaining significant ownership over your tooling.
• World‑Class Facilities: Validate your automated benchmarks against physical, full‑scale dual‑arm robotic testbeds and hardware‑in‑the‑loop space flight simulators.

Additional Position and Application Info

This posting is for an existing vacancy. Thank you to all who apply. Candidates who are successfully selected for an interview will be notified by email.

We encourage all qualified applicants to apply, even if they don't check every single box. We recognize the impact of imposter syndrome and value practical capability, initiative, and drive.

If the posted salary range, work location, or anything else does not exactly align with your expectations, please let us know. We'd hate to miss out on the opportunity of working together without a discussion first.

Equity

We are an equal opportunity employer committed to building a diverse and inclusive workplace. We welcome applications from all qualified individuals, including women, Indigenous peoples, persons with disabilities, members of visible minorities, and people of all sexual orientations and gender identities.

Accommodation

If you require accommodation during any stage of the recruitment process, please contact us at [e-mail] , and we will work with you to meet your needs.