**About Agrex.ai**
Agrex.ai is a pioneering video analytics company dedicated to transforming existing CCTV infrastructure into intelligent, real-time monitoring systems across diverse industries such as retail, manufacturing, banking, logistics, and education. Our innovative platform offers comprehensive analytics including people and vehicle detection, SOP/compliance monitoring, and operational insights. These solutions are deployed efficiently on edge devices like NVIDIA Jetson and in cloud environments, enabling actionable intelligence and enhanced decision-making capabilities.
—
**Position: Computer Vision Intern**
We are seeking a motivated and passionate Computer Vision Intern to join our dynamic team. This role offers a unique opportunity to contribute to cutting-edge video analytics applications and gain practical experience with state-of-the-art machine learning models, video processing pipelines, and deployment on specialized hardware such as NVIDIA Jetson and GPUs. The intern will collaborate closely with our Product and Engineering teams to deliver impactful solutions addressing real-world challenges.
—
**Key Responsibilities**
– Collaborate in the implementation, training, and evaluation of computer vision models focused on detection, segmentation, and tracking, leveraging architectures such as YOLO, RTDETR, Mask R-CNN, ByteTrack, and OC-SORT.
– Assist in integrating these models into NVIDIA DeepStream or GStreamer pipelines; facilitate model exportation to ONNX and TensorRT formats and optimize for enhanced speed and efficiency.
– Contribute to developing features like people and vehicle counting, dwell time and queue analytics, intrusion detection, personal protective equipment (PPE) compliance checks, and fundamental re-identification (re-ID) tasks.
– Work hands-on with various video stream formats including RTSP, H.264, and H.265, managing frame sampling strategies, mitigating jitter effects, and implementing robust reconnection logic.
– Write modular, maintainable Python code—utilizing libraries such as NumPy, OpenCV, and PyTorch/TensorFlow—with occasional use of C++ for performance-critical components; assist in unit testing and benchmarking processes.
– Support data preparation efforts by curating datasets, establishing labeling guidelines, performing annotations using tools like CVAT or Label Studio, and applying basic data augmentation techniques.
– Systematically track experimental results using relevant metrics including mean Average Precision (mAP), recall rates, frames per second (FPS), and GPU utilization; document findings and insights to inform future development.
– Participate in diagnosing and resolving production issues, contributing to prototype development and rapid problem-solving.
– Adhere to best practices for securing and managing enterprise-grade video data, ensuring data privacy and compliance.
—
**Candidate Profile**
**Required Skills and Knowledge:**
– Proficiency in Python programming with experience in libraries such as NumPy, OpenCV, and deep learning frameworks like PyTorch or TensorFlow.
– Solid understanding of fundamental computer vision and machine learning concepts including convolutional operations, Intersection over Union (IoU), Non-Maximum Suppression (NMS), data augmentation methods, overfitting avoidance, and evaluation metrics.
– Hands-on exposure to at least one of the following domains: object detection, segmentation, or tracking.
– Ability to independently learn from technical documentation, research articles, and enthusiastically develop small utilities or tools to support project goals.
**Desirable Qualifications (Bonus):**
– Familiarity with the NVIDIA technology stack such as CUDA programming, TensorRT optimization, and DeepStream SDK for video analytics.
– Experience working with NVIDIA Jetson platforms (e.g., Xavier, Orin) or other edge computing devices.
– Knowledge of multimedia frameworks and codecs, including GStreamer and FFmpeg.
– Exposure to machine learning lifecycle tools such as Data Version Control (DVC), Weights & Biases (W&B), Docker for containerization, and building automated ML pipelines.
– Interest or prior experience in advanced topics like re-identification (re-ID), human pose estimation, Optical Character Recognition (OCR), or multi-camera system integration.
– Previous internships or project work involving video analytics in domains like retail, manufacturing, warehouses, or surveillance would be a strong advantage.
—
**Educational Qualifications**
Candidates pursuing or holding a Bachelor’s or Master’s degree in Computer Science, Electrical Engineering, Mathematics, or related fields—or possessing equivalent practical experience—are encouraged to apply.
—
**Why Join Agrex.ai?**
At Agrex.ai, you will immerse yourself in a fast-paced, innovative environment where you can sharpen your technical skills, collaborate with experienced professionals, and contribute to impactful products that redefine video analytics across multiple industries. This internship is an excellent opportunity to gain hands-on experience with machine learning in production contexts, deepen your understanding of edge and cloud-based video processing, and build expertise in AI-powered operational intelligence.
—
We look forward to welcoming enthusiastic learners who are eager to make significant contributions and grow alongside us in shaping the future of intelligent video analytics.
