Computer Vision Engineer

Computer Vision Engineer – DeepStream / Edge AI

About the Role

We are looking for a hands-on Computer Vision Engineer to design, develop, and deploy real-time video analytics and visual perception systems.

The role involves building production-grade AI pipelines using NVIDIA DeepStream, TensorRT, and modern Computer Vision models running on GPU and edge platforms.

The selected candidate will work on advanced maritime and surveillance applications involving multi-camera analytics, object detection, tracking, landmark recognition, visual navigation, and real-time situational awareness.

Key Responsibilities

Computer Vision Development

• Develop and deploy real-time object detection, tracking, and classification systems.

• Build video analytics pipelines for EO/IR camera systems.

• Develop visual landmark detection and recognition algorithms.

• Implement vessel, object, and target tracking solutions.

• Design image enhancement and video preprocessing pipelines.

AI Model Development

• Train, fine-tune, and evaluate deep learning models using PyTorch.

• Work with YOLO, RT-DETR, segmentation, classification, and tracking models.

• Create and manage datasets, annotations, and evaluation benchmarks.

• Improve model performance under varying weather, lighting, and environmental conditions.

Real-Time Video Analytics

• Build and maintain NVIDIA DeepStream pipelines.

• Integrate RTSP/IP camera streams and multi-camera systems.

• Implement tracking algorithms such as ByteTrack, DeepSORT, and NvDCF.

• Develop event-driven analytics and alerting systems.

Optimization & Deployment

• Optimize inference using TensorRT, ONNX, CUDA, and GPU acceleration.

• Deploy applications on NVIDIA Jetson platforms and GPU servers.

• Improve latency, throughput, and resource utilization.

• Support field deployments, testing, and troubleshooting.

System Integration

• Integrate AI modules with software applications, APIs, databases, and external sensors.

• Collaborate with software, deployment, and hardware teams to deliver end-to-end solutions.

• Maintain documentation, benchmarks, and deployment procedures.

Required Technical Skills

Programming

• Python

• Linux development environment

• Bash scripting

Computer Vision & AI

• OpenCV

• PyTorch

• Object Detection

• Multi-Object Tracking

• Image Processing

• Video Analytics

Deep Learning Models

• YOLO (v8/v11 preferred)

• Classification Models

• Segmentation Models

Deployment & Optimization

• NVIDIA DeepStream

• TensorRT

• ONNX Runtime

• CUDA

• Docker

Video Streaming

• RTSP

• GStreamer

• Multi-camera processing

Preferred Skills

• NVIDIA Jetson (Orin / Xavier)

• DeepSORT / ByteTrack / NvDCF

• Vision-Language Models (Qwen, MiniCPM, LLaVA)

• OCR and Document Vision

• GIS / Geospatial Visualization

• Maritime, Defence, Surveillance, or Smart City domain experience

• Sensor Fusion (Camera + GPS / Radar / AIS)

Desired Candidate Profile

• Strong problem-solving and debugging skills.

• Experience handling production deployments and field trials.

• Ability to optimize systems for latency, throughput, and reliability.

• Comfortable working across AI, software, and deployment domains.

• Self-driven engineer capable of taking ownership from development to deployment.

Qualification

• B.Tech / M.Tech in Computer Science, AI, Electronics, Robotics, Mechatronics, or related discipline.

• 3–6 years of relevant Computer Vision development experience.