Visual Grab is looking for a motivated Computer Vision Intern who wants to work on real-world AI and Computer Vision problems across manufacturing, robotics, transportation, healthcare, retail, security, agriculture, and other industries.

The intern will work closely with our Computer Vision team on the complete development lifecycle—from understanding a problem and preparing datasets to training, evaluating, optimizing, and deploying Computer Vision models.

Key Responsibilities

- Develop and experiment with Computer Vision and Deep Learning models.
- Work on Object Detection, Classification, Segmentation, Pose Estimation, OCR, Multi-Object Tracking, Activity Recognition, Anomaly Detection, Visual Inspection, and 3D Vision.
- Prepare, clean, annotate, and analyze image/video datasets.
- Train and fine-tune models using frameworks such as PyTorch, TensorFlow, Ultralytics YOLO, OpenCV, and Hugging Face.
- Evaluate models using appropriate metrics such as Precision, Recall, F1-score, IoU, mAP, latency, and FPS.
- Improve model accuracy and robustness through data augmentation, hyperparameter tuning, error analysis, and dataset improvement.
- Optimize models for real-time and edge deployment using techniques such as quantization, pruning, ONNX and TensorRT.
- Build Computer Vision pipelines integrating cameras, models, APIs, databases, and application software.
- Work on research-oriented Proof-of-Concepts and customer projects.
- Explore Vision Transformers, Vision-Language Models, Foundation Models, Multimodal AI and Agentic AI.
- Document experiments, results, observations, and technical findings.

Desired Skills Required

- Python programming
- Basic understanding of Machine Learning and Deep Learning
- Fundamentals of Computer Vision
- OpenCV
- NumPy




- Understanding of CNNs and neural networks
- Willingness to experiment, research, and learn

Positive to Have

- PyTorch or TensorFlow
- YOLO
- Image segmentation models
- Transformers / Vision Transformers
- Tracking algorithms such as ByteTrack, DeepSORT or BoT-SORT
- Git/GitHub
- Linux
- Docker
- ONNX / TensorRT
- NVIDIA GPU development
- Raspberry Pi / Jetson or other edge devices

Example Projects Interns may get exposure to projects such as:
- Automated Visual Inspection for Manufacturing
- Defect and Anomaly Detection
- Worker–Machine Safety Analytics
- Object Detection and Tracking
- Human Pose and Activity Recognition
- Retail and Logistics Video Analytics
- Traffic and Transportation Analytics
- Robotics Vision and Object Picking
- OCR and Document Intelligence
- Multi-Camera Tracking
- 3D Vision and Depth Estimation
- Edge AI deployment

Who Should Apply? Students or recent graduates from:

Computer Science | AI/ML | Electronics | Robotics | Mathematics | Data Science | related engineering disciplines

We are particularly interested in candidates who enjoy building things, experimenting with models, reading research papers, debugging difficult problems, and converting research ideas into working products.

What You Will Gain This internship is designed to provide exposure beyond simply training models.

You will learn how real Computer Vision products are developed through:

Problem Definition → Data Collection → Annotation → Model Selection → Training → Evaluation → Error Analysis → Optimization → Deployment → Real-World Validation

You will work with a team focused specifically on Computer Vision AI and real-world product development.

Business Unit: Visual Grab – Computer Vision

- Company: Miracle Eye Private Limited

📌 Computer Vision Intern (Dehradun)
🏢 Computer Vision AI Visual Grab
📍 Dehradun