onsite
Software Engineer - Computer Vision
Software Engineer (Computer Vision)
The Computer Vision Software Engineer will develop and refine AI and machine learning models for advanced analytics, including features like people and vehicle detection, license plate recognition, and real-time facial recognition. This role involves implementing deep learning networks using frameworks like TensorFlow and PyTorch, and writing clean C++ code for computer vision algorithms.
About the role
About the Role
At the forefront of innovation, the Computer Vision team develops the artificial intelligence and machine learning models that power Verkada's advanced analytics. Their responsibilities include creating and refining algorithms for features such as people and vehicle detection, license plate recognition, and other intelligent video analysis tools. Recent projects for this role include:
- Implementing and deploying a binary classifier using TensorFlow for detecting the binary states across millions of cameras
- Detecting unusual object addition/removal in a scene
- Detecting and counting object and people frequencies
- Training text image embedding models and vision language models
- Training license plate recognition models and implementing LPR on the edge
- Training facial recognition models and implementing real-time facial recognition
What You'll Do
- C++ - writing clean, modular, C++ code
- Traditional computer vision algorithms
- Training deep learning networks using PyTorch, TensorFlow, Keras, or similar
- Data structures and architecture
- Must be willing and able to work onsite five days per week
What You Bring
- Bachelor's Degree in Computer Science, preferably with research experience
- 2+ years of industry software engineering experience
- 1+ years of work or research experience with current neural net frameworks
- Mastery of at least one practical programming language
- Experience working in an agile team software development environment
Skills
Artificial IntelligenceMachine Learningpeople detectionvehicle detectionlicense plate recognitionintelligent video analysisbinary classifierTensorFlowobject detectionpeople countingtext image embedding modelsvision language modelsfacial recognition modelsC++traditional computer vision algorithmsdeep learning networksPyTorchKerasData Structuresarchitecture