Machine Learning Engineer

About the role

This position involves leading the development and scaling of biometric face recognition systems. You will be responsible for building and refining machine learning models and managing the entire lifecycle of ML systems on AWS. The specific level of this role will be determined after the interview stages.

Key facts

What you'll do

• Drive the creation of computer vision solutions for biometric applications, covering face attributes, detection, quality assessment, and recognition.
• Conduct thorough analysis of fairness and performance for biometric models across diverse datasets and conditions.
• Design, train, and optimize models using PyTorch, Tensorflow, or JAX.
• Oversee and advance complete machine learning pipelines, from data acquisition to deployment, including creating automated workflows for data processing.
• Manage the path to production for ML models, optimizing them for fast inference and deploying them on AWS.
• Provide guidance to other ML engineers, review code and designs, and promote engineering best practices within the team.

Requirements

• A minimum of 5 years of professional experience in machine learning, with at least 3 years focused on biometrics or face analysis.
• Extensive knowledge in computer vision and biometrics, particularly face recognition.
• Understanding of algorithmic bias in computer vision, with practical experience in measuring and reducing its impact.
• Strong Python programming skills, including proficiency with machine learning and computer vision libraries like Pillow, OpenCV, PyTorch, and others, writing production-ready code.
• Experience in designing end-to-end ML pipelines and using workflow orchestrators such as Airflow.
• Hands-on experience with scaling training on multi-GPU systems and deploying services on AWS, including SageMaker, EC2, and EKS.

Nice to have

• Published research in top-tier computer vision conferences like CVPR, ICCV, ECCV, or FG, related to face recognition, image quality, or fairness.
• Experience with vector databases like Milvus or Faiss, and approximate nearest neighbor search.
• Familiarity with privacy, security, and compliance aspects of biometric systems.
• Experience adapting models for mobile or edge devices using frameworks like CoreML, LiteRT, or TFLite.
• Experience generating synthetic faces for training using GANs or diffusion models.
• Excellent communication abilities.

Skills & tools

• Python
• PyTorch
• Tensorflow
• JAX
• AWS (SageMaker, EC2, EKS)
• Airflow
• Computer Vision libraries (OpenCV, Pillow)
• Biometrics
• Face Recognition
• Machine Learning Pipelines
• Model Optimization (quantization, distillation, TensorRT/ONNX)

Practical notes

• Jumio is committed to diversity and inclusion, welcoming applications from all backgrounds.
• Jumio is a technology company focused on preventing online identity fraud and financial crime.
• Personal information will be used solely for the application and recruitment process, as detailed in Jumio's Applicant Privacy Notice.