Senior Computer Vision Engineer – Vision Systems & Inference

Experience: 5+ years

Role Overview

We are looking for an experienced Computer Vision Engineer who has built and deployed real-world computer vision analytics systems involving object detection, segmentation, tracking, OCR, video analytics, and multi-camera processing.

The ideal candidate should have strong computer vision fundamentals along with hands-on experience designing inference architectures, high-throughput data pipelines, model integration, system optimization, and production deployment.

This is primarily an applied computer vision engineering role, with emphasis on building reliable and scalable vision systems.

Key ResponsibilitiesComputer Vision & Model Implementation

· Implement and fine-tune computer vision models for object detection, segmentation, tracking, OCR, and video analytics.

· Work extensively with architectures such as YOLO and other modern detection/segmentation models.

· Integrate and adapt pretrained models for application-specific datasets and requirements.

· Evaluate models using appropriate accuracy, robustness, and performance metrics.

Multi-Camera & Object Tracking

· Design and implement multi-object and multi-camera tracking systems.

· Work with tracking approaches such as ByteTrack, DeepSORT, ReID-based tracking, or similar techniques.

· Handle object association, identity consistency, camera streams, and temporal information in real-world environments.

OCR & Video Analytics

· Work with modern OCR pipelines for applications such as ANPR, text detection, and recognition.

· Build complete video-processing workflows combining detection, tracking, OCR, segmentation, and business analytics.

Inference Architecture & Performance Optimization

· Design productive inference pipelines for real-time and high-throughput video processing.





· Optimize GPU/CPU utilization, model execution, preprocessing, post-processing, and data movement.

· Use multiprocessing, multithreading, asynchronous processing, queues, batching, and pipeline parallelism where appropriate.

· Identify and resolve inference bottlenecks involving compute, memory, I/O, or system architecture.

Data Pipeline Engineering

· Build scalable pipelines capable of processing large image/video datasets and multiple camera streams.

· Develop efficient preprocessing, augmentation, validation, storage, and data-ingestion workflows.

· Design systems that can reliably process continuous or large-volume visual data.

Deployment

· Deploy computer vision applications across GPU servers, cloud infrastructure, and edge platforms such as NVIDIA Jetson.

· Optimize models using approaches such as TensorRT, ONNX, FP16/INT8 inference, quantization, and batching where applicable.

· Build maintainable production systems around trained computer vision models.

Computer Vision Fundamentals

· Apply strong knowledge of both classical computer vision and deep-learning-based vision techniques.

· Understand image processing, geometric transformations, camera systems, feature extraction, filtering, image enhancement, and related concepts.

· Understand model training, transfer learning, loss functions, optimization, hyperparameter tuning, and model evaluation.

Required Skills

· 5+ years of hands-on experience in Computer Vision / Machine Learning.





· Strong experience with Python, OpenCV, PyTorch and/or TensorFlow.

· Production experience with YOLO or similar object detection and segmentation architectures.

· Strong experience with object tracking and video analytics systems.

· Experience building or integrating OCR pipelines.

· Experience handling multiple camera/video streams.

· Strong understanding of inference architecture and performance optimization.

· Experience with multiprocessing, multithreading, asynchronous pipelines, and GPU-based inference.

· Experience building pipelines for large-scale image/video datasets.

· Good understanding of computer vision fundamentals and deep-learning concepts.

· Experience fine-tuning and adapting pretrained computer vision models.

Good to Have

· Experience with multi-camera multi-object tracking (MTMC) and ReID systems.

· Experience with segmentation architectures such as U-Net, U-Net++, SegFormer, or transformer-based vision models.

· Experience with depth maps, stereo vision, point clouds, or other 3D computer vision data.

· Experience with TensorRT, ONNX Runtime, CUDA, Triton Inference Server, or similar inference technologies.

· Experience deploying models on NVIDIA Jetson or other edge devices.

· Exposure to distributed processing, streaming architectures, RTSP/CCTV systems, or large-scale video analytics platforms.

· Experience profiling and optimizing end-to-end computer vision applications.

Education

B.Tech / M.Tech in Computer Science, Computer Engineering, Information Technology, Electronics, Electrical Engineering, or a related field from a reputed institution, or equivalent relevant industry experience.

[email protected]

Experience:

- Computer vision: 6 years (Preferred)

Work Location: In person

📌 Computer Vision Engineer (Navi Mumbai)
🏢 Blue Ocean Systems Infotech
📍 Navi Mumbai