What You'll Do

Design and develop real-time computer vision pipelines for person and object detection.
Build pose estimation and movement analysis models to interpret posture, gestures, and motion.
Develop multi-object tracking systems with identity persistence across video frames.
Optimize AI models for low-latency inference on edge and near-edge devices.
Ensure reliable performance across varying lighting conditions, camera angles, and occlusions.
Integrate vision models with live video streaming and processing frameworks.
Evaluate and benchmark models using real-world datasets and live video feeds.
Collaborate with cross-functional teams to deliver scalable AI solutions.
Continuously improve detection accuracy, tracking stability, and inference performance.

Required SkillsComputer Vision & Deep Learning

Object Detection: YOLO, Faster R-CNN, SSD
Pose Estimation: OpenPose, HRNet, MediaPipe
Multi-Object Tracking: DeepSORT, ByteTrack, OC-SORT

Programming & Frameworks

Python
PyTorch or TensorFlow
OpenCV
GStreamer / FFmpeg





Additional Expertise

Image Processing & Feature Extraction
GPU Acceleration & Performance Optimization
TensorRT, ONNX, Quantization
Real-time Video Processing
Model Evaluation (Precision, Recall, Tracking Accuracy, Latency)

Experience Required

5+ years of hands-on experience in Computer Vision or Applied AI.
Experience deploying real-time vision systems in production.
Robust exposure to live video feeds, real-world datasets, and challenging deployment environments.

What You'll Build

You'll Contribute To Next-generation AI-powered Vision Systems Capable Of

Real-time human detection and tracking
Pose and movement analysis
Behavioral analytics
Intelligent video understanding
Scalable AI solutions for real-world applications

Skills: multi-object tracking (deepsort, bytetrack, oc-sort),computer vision (object detection & pose estimation),python (pytorch/tensorflow)

📌 Computer Vision Engineer – Detection (Bengaluru)
🏢 Durapid Technologies Private
📍 Bengaluru