Machine Learning Engineer

What You’ll Do (Essential Responsibilities):

• Build, train, and ship ML models for identity verification use cases such as biometric matching, liveness / anti-spoofing, identity document processing (OCR/extraction), and fraud detection (team assignment based on experience). 
• Prepare large, noisy datasets: ingestion, validation, cleaning, deduplication, labeling strategy, and dataset QA to improve model performance and reliability. 
• Design experiments, evaluation protocols, and success metrics (offline and online), iterate based on measurable business impact (detection rates, fraud losses, false positives). 
• Develop production-grade training and inference pipelines on AWS with strong reproducibility, monitoring, and cost controls. 
• Productionize models as resilient services and libraries in Python; collaborate with platform teams on APIs, latency and observability. 
• Contribute to the transformation of our IDV engine: modernizing legacy components, improving modularity, and raising quality, performance, and maintainability. 
• Work closely with Product, Customer Success, and Platform Engineering teams to ensure ML solutions meet privacy, compliance, and reliability requirements. 
• Support other engineers through design reviews, code reviews, and knowledge sharing; help raise the technical bar across the team. 

Who You Are (Soft Skills/Attributes):

• Analytical, creative, and innovative. You eagerly try new things and learn from your experiences. 
• Logical and creative problem-solver with the ability to summarize issues effectively. 
• Excellent time manager, with the ability to efficiently shift priorities. 
• Clear and concise communicator. You effectively set expectations and raise issues as needed. 
• Strong team player with a positive attitude and ability to adapt to changes. 
• Demonstrate ability to work with ambiguous requirements, adapt, and learn. 
• Knowledgeable with broad technical experience in all design phases.  
• Competent in programming and debugging across multiple modules and dealing with related external dependencies. 
• Willing to learn and adapt to new technologies. 

What You Need (Required Knowledge, Skills & Abilities):

• Bachelors degree in computer science or related field paired with knowledge, skills and abilities typically gained from 2-5 years of experience in applied machine learning / ML engineering with strong software engineering fundamentals (or equivalent combination of education and experience). 
• Strong Python skills and experience building production ready code.  
• Demonstrated experience solving computer vision tasks with ML models utilizing PyTorch or Tensorflow. 
• Strong computer vision background, including experience with CNNs, vision transformers, and foundation models. 
• Proven ability to work with large datasets and build reliable data preprocessing/feature engineering pipelines; comfort with distributed data tooling is a plus. 
• Clear communication skills and the ability to work effectively across engineering, product, and operations stakeholders. 

What Would be Nice (Preferred Skills & Experience):

• Experience running ML in production: containerization (Docker), CI/CD, monitoring, and model/version management; ability to troubleshoot data/model issues end-to-end. 
• Experience optimizing models for real-time constraints (quantization, distillation, pruning, ONNX) and performance tuning for CPU/GPU inference. 
• Model understanding / interpretability experience (e.g., Grad-CAM, saliency maps, error slicing, and targeted evaluation). 
• Experience with experiment tracking (e.g., MLflow, Weights & Biases) and strong habits around reproducibility. 

Our Tech Stack Includes:

• Cloud: AWS (AWS-native services for AI/ML and production workloads) 
• Languages: Python 
• Data & Storage: S3, DynamoDB, MongoDB (varies by service) 
• ML Platform: SageMaker (plus standard tooling for training, evaluation, and monitoring) 
• ML Tools: Tensorflow, PyTorch, Matplotlib, Pandas, Scikit-learn, OpenCV, Pillow 
• Deployment: Containers and orchestration (ECS/EKS), CI/CD, observability