Machine Learning Scientist

About DarkVision Technologies Inc.

DarkVision is a technology company based in Vancouver, established in 2013, pioneering advancements in the industrial imaging sector. We have developed the most sophisticated acoustic-based imaging platform available, transforming how clients assess and visualize the condition of critical industrial assets.

Supported by Blackstone, a leading global private equity firm, DarkVision is rapidly growing its diverse team of Software, Mechanical, and Machine Learning Engineers along with Data Analysts to support current and upcoming product demands.

Our employees engage with cutting-edge technology that bridges scientific innovation and tangible applications. Join us on our mission to become the worldwide leader in industrial imaging.

Your Role

We are looking for a Machine Learning Scientist for our Imaging & AI group. The selected candidate will research, craft, and prototype deep learning frameworks that fuel our automated analytical instruments. Your work will tackle complex computer vision challenges aimed at enabling scalable interpretation of industrial data.

Our ultrasound imaging system amasses petabyte-scale datasets, identifying minute defects in extensive industrial assets. Developing algorithms to automate data processing, classification, and visualization is critical. You will confront mathematical and architectural problems to build models that efficiently learn representations tailored for real-world industrial scenarios.

This position requires working from our North Vancouver headquarters, where amenities include a gym, squash court, steam room, and climbing wall.

Team Environment

As part of the Imaging & AI team, you will collaborate with scientists and engineers across disciplines. This team focuses on early-stage ideation, research, experimentation, and development, ensuring seamless integration of models into product offerings through close technical collaboration.

Key Responsibilities

• Conduct applied research and engineer cutting-edge deep learning architectures targeted at large-scale industrial challenges such as object detection, instance and semantic segmentation, and anomaly detection.
• Rapidly prototype and validate model concepts using Python and deep learning tools, managing experimental phases to verify statistical significance and efficacy prior to engineering integration.
• Collaborate with data analysts and scientists to explore, organize, and structure datasets while defining extraction and preprocessing methodologies for robust training.
• Monitor and enhance model performance continuously by researching measures to improve accuracy and resilience on complex real-world datasets.
• Document model architectures, training protocols, and dataset handling processes to support reproducibility and facilitate engineering handoff.

Preferred Qualifications

• Master’s degree in Computer Science, Electrical Engineering, or related disciplines; a PhD is favored.
• Over three years of experience designing or researching deep learning models.
• Strong skills in Python and familiarity with deep learning frameworks like PyTorch and JAX.
• Capability to design rigorous experiments and analyze complex results.

Additional Advantageous Skills

• Experience handling medical or industrial ultrasound image data.
• Publications in premier computer vision conferences such as CVPR or ICCV.
• Knowledge of self-supervised or unsupervised learning techniques.
• Practical experience with ML lifecycle tools (e.g., MLFlow, Amazon SageMaker, GPC Vertex AI).
• Experience with orchestration platforms such as Kubeflow, Prefect, or Airflow.
• Leadership or mentoring experience within research or scientific contexts.
• Ability to thrive in a dynamic and autonomous work environment.
• Excellent skills in communicating research findings and producing detailed technical reports.

Compensation and Benefits

The anticipated salary range for this position is $100,000 to $150,000 annually. This role includes eligibility for variable compensation, potentially awarded as bonuses or alternate forms.