Babel Street
Software Engineer
About this role
Babel Street seeks an early-career Software Engineer to join the Image & Computer Vision AI team, supporting the development and deployment of computer vision capabilities for identity intelligence and risk operations. You'll work on image search, object detection, facial matching, and multimodal AI features while collaborating with senior engineers in a hybrid environment.
What you'll do
- Develop and deploy computer vision systems for image analysis and visual intelligence applications
- Build image search, object detection, scene understanding, and facial matching workflows
- Integrate computer vision models with multimodal LLM systems for natural language image reasoning
- Support geolocation and multimodal intelligence feature development
- Test, validate, and improve reliability of vision capabilities in production
- Collaborate with senior engineers and cross-functional teams on implementation and iteration
What they're looking for
- Computer vision and image processing
- Machine learning and applied AI
- Python or similar programming languages
- Deep learning frameworks (TensorFlow, PyTorch, etc.)
- Object detection and facial recognition techniques
- Multimodal AI systems and LLM integration
- Software engineering practices and testing
- Geolocation and spatial data analysis
Opens the official application on the employer’s site. No login required.
Babel Street
Babel Street develops intelligence platforms that extract, match, and analyze complex data from diverse sources using advanced AI techniques including natural language processing, computer vision, and web data harvesting. The company is hiring software engineers across multiple specializations—from data infrastructure and NLP to computer vision—to build and maintain the core systems powering their identity intelligence and data analytics capabilities.
- Website
- babelstreet.com
Likely interview questions
- Walk us through a computer vision or image processing project you've worked on—what was the problem and how did you approach it?
- How would you approach building a facial matching system that needs to balance accuracy with performance at scale?