Machine Learning Engineer
JumioCanada2w ago
Machine LearningComputer VisionEngineeringremotecurated-jd
Job description
Machine Learning Engineer at Jumio.
About the role
Jumio is seeking a Senior or Staff Machine Learning Engineer to focus on computer vision and biometrics. You will lead the development and scaling of face recognition systems, managing the entire lifecycle from model training to production deployment on AWS. Your final job title will be determined based on your interview performance.
Key facts
What you'll do
- Design and develop computer vision systems for biometric tasks including detection, quality assessment, and face recognition.
- Conduct thorough fairness benchmarking and analysis across diverse datasets to ensure model equity.
- Architect, train, and refine models using JAX, Tensorflow, or PyTorch.
- Manage end-to-end ML pipelines, including data cleaning and ingestion via Airflow.
- Curate balanced training sets and generate synthetic data to address diversity and quality requirements.
- Optimize models for low-latency inference using distillation, quantization, ONNX, or TensorRT.
- Oversee deployment on AWS and provide mentorship to other engineers while enforcing technical standards.
Requirements
- 5+ years of professional experience in Machine Learning.
- 3+ years of specific experience in Face Analysis or Biometrics.
- Proficiency in Python and vision libraries such as OpenCV, Pillow, and PyTorch.
- Experience with end-to-end ML pipeline architecture and workflow orchestration tools like Airflow.
- Practical knowledge of measuring and mitigating algorithmic bias in computer vision.
- Hands-on experience scaling training on multi-GPU clusters and deploying on AWS services like EKS, EC2, and SageMaker.
Nice to have
- Research publications in conferences such as ICCV, CVPR, ECCV, or FG.
- Experience with approximate nearest neighbor algorithms and vector databases like Faiss or Milvus.
- Knowledge of security, privacy, and compliance standards for biometrics.
- Experience porting models to mobile or edge devices using TFLite, LiteRT, or CoreML.
- Background in using diffusion models or GANs for synthetic data generation.
Skills & tools
- Python, PyTorch, Tensorflow, JAX
- AWS (SageMaker, EKS, EC2)
- Airflow
- OpenCV, Pillow
- TensorRT, ONNX
- Quantization, Distillation
Practical notes
Jumio values integrity, diversity, empowerment, accountability, and innovation. The company is an equal opportunity employer. Personal data submitted during the application process is handled according to the Jumio Applicant Privacy Notice.