Computer Vision Developer AI Attendance & Surveillance

About the Role

We're building the next generation of AI-powered attendance and security systems - and we need a Computer Vision Developer to lead that effort. You'll work on face recognition for contactless attendance, AI-driven CCTV analytics (intrusion detection, people counting, ANPR), and edge-deployed inference models that run on our hardware devices.

This is a hands-on, research-to-production role. You'll prototype in Python, train models, optimize them for edge deployment, and ship features that run on thousands of devices in the field.

Key Responsibilities

• Develop and optimize face recognition pipelines for real-time attendance - detection, alignment, embedding extraction, and matching against enrolled databases.
• Build AI-powered video analytics modules: intrusion detection, loitering alerts, people counting, vehicle ANPR (Automatic Number Plate Recognition), and occupancy monitoring.
• Train, fine-tune, and evaluate object detection and classification models (YOLO, SSD, EfficientDet) on custom datasets for attendance and surveillance use cases.
• Optimize trained models for edge deployment - quantization, pruning, and conversion to ONNX/TensorRT/TFLite for inference on ARM boards, NVIDIA Jetson, or custom hardware.
• Build data pipelines for collecting, labeling, augmenting, and versioning training datasets.
• Develop REST APIs and microservices to serve CV inference results to our attendance and access control software stack.
• Collaborate with the hardware team to integrate camera feeds, IR sensors, and depth cameras into the CV pipeline.
• Stay current with SOTA research in face recognition, liveness detection (anti-spoofing), and video understanding.

Required Qualifications

• 2-6 years of professional experience in computer vision or applied deep learning.
• Strong Python skills - production-grade, not just notebook-level prototyping.
• Hands-on experience with OpenCV, PyTorch or TensorFlow, and at least one detection framework (YOLO, Detectron2, MMDetection).
• Working knowledge of face recognition techniques - ArcFace, FaceNet, InsightFace, or similar embedding-based approaches.
• Experience deploying CV models on edge devices or resource-constrained hardware.
• Understanding of camera systems - resolution, frame rate, IR illumination, lens distortion, and their impact on model performance.
• Familiarity with Linux, Docker, and Git.

Preferred (Nice-to-Have)

• Experience with liveness detection / anti-spoofing (2D photo attacks, video replay, 3D mask detection).
• Exposure to NVIDIA Jetson (Nano/Xavier/Orin) or Rockchip edge platforms.
• Experience with ANPR or OCR pipelines.
• Familiarity with Ultralytics YOLO training and export workflows.
• Published papers or open-source contributions in CV/ML.

Disclaimer: This job posting has been aggregated from external source. Role details, content, and availability are subject to change. Applicants are advised to confirm the latest information directly on the company website before applying.