This role is based in-person at our Vancouver office.

In this role, you will apply machine learning and computer vision techniques to real-world challenges like robotic part picking and localization in structured, high-speed applications. You’ll prototype, evaluate, and improve models that are deployed on factory floors in industries such as automotive and industrial manufacturing.

Employee Value Proposition (EVP)

• Purpose : You’ll contribute to the intelligence behind robotic systems that perform precise, high-speed automation tasks such as part picking and placement for stamped metal components or machined assemblies.
• Growth: You’ll gain hands-on experience applying academic concepts to production workflows and working with internal datasets, building robust models, and learning from system behavior in real deployments.
• Motivators: You’ll be part of a collaborative, fast-moving team, and see your models tested in simulation and on real industrial robots used in customer-facing solutions.

Major Objectives

• Prototype and Evaluate Vision Models Within the first 90 days, implement machine learning models for object detection, depth estimation, or 6-DoF pose estimation. Benchmark performance using internal datasets that reflect real manufacturing conditions. Tools: PyTorch, internal GPU cluster, dataset tools- Translate Research into Production-Relevant Improvements Identify and prototype methods from recent ML or computer vision research. Adapt them to our application domain and evaluate them against production baselines. Document findings and trade-offs. Focus: Model speed, stability, accuracy under varying lighting and part geometry- Enhance Synthetic Data Generation for Model Training Contribute improvements to the synthetic data generation pipeline, focusing on expanding variation in object shape, material, and pose. Help ensure the dataset supports model generalization across production use cases.

Critical Subtasks

• Evaluate the ML Development Environment In your first month, review the current training and validation tools. Identify areas for performance or usability improvements and contribute one concrete change by mid-term.
• Collaborate Cross-Functionally and Debug Model Issues Work with robotics and software engineers to understand deployment requirements and constraints. Assist in diagnosing issues with model performance observed during robotic testing or simulation, and help implement fixes or improvements.
• Own and Deliver a Scoped ML Project Lead a focused initiative such as testing a new augmentation strategy, developing a lightweight evaluation tool, or experimenting with model modifications for improved robustness. Present outcomes with metrics and insights at the end of the term.
• Support Research on a Strategic Vision Problem Join early investigations into longer-term capabilities, such as handling part occlusion or improving model behavior with similar-looking parts. Conduct benchmarking and literature review to inform future roadmap decisions.

Culture and Situation Fit

You’ll thrive if you value initiative, technical depth, and seeing data as a design lever, not just input. Apera AI is fast-paced, collaborative, and impact-driven. Engineers here build systems that make AI dependable in messy, real-world conditions

You’ll Thrive Here If You

• Want to apply ML to real-world problems in industrial automation.
• Are excited to see your work influence how robotic systems are built and deployed.
• Enjoy solving practical problems with research-informed tools.

Qualifications

• Proficiency in Python and machine learning frameworks (e.g., PyTorch).
• Understanding of computer vision fundamentals (e.g., detection, segmentation, 3D geometry).
• Familiarity with model training, tuning, and evaluation workflows.
• Interest in robotics or applying ML in production-grade software.

Bonus Experience (Not Required)

• Experience with synthetic data generation or tools like Blender.
• Exposure to 6-DoF pose estimation, point cloud processing, or depth sensing.
• Experience working in Linux or Docker-based environments.
• Familiarity with AWS services used in ML development workflows (e.g., S3, EC2, SageMaker).

To apply: Please ensure you upload both your resume and transcript, either combined into a single file or as separate files.

• Application Questions: Share how you think and solve problems through ten short questions, no memorization or trick questions.
• Assessment: After applying, you will be selected to complete a TestGorilla assessment and walk away with results you can reuse in any future job applications.
• Interviews: Take part in two one-hour conversations focused on real problem-solving, how you work, and what you want to learn from the co-op.
• Background Check: A brief check is completed as required before an offer.
• Timeline: Receive a decision within three to four weeks, so you can plan your journey with confidence.
• Human Review: Your application is reviewed by real engineers and people-team members who care about your growth.