Computer Vision Developer Internship at IIT GUWAHATI… · IIT…

Location: IIT Guwahati Campus, Assam
Project Link: https://iitgdrishti.in/

About the Role

We are looking for a Computer Vision Developer to build, train, optimise, and deploy AI models for image and video applications. The role involves GPU-based model development and efficient deployment on edge processors, with Python for AI/ML development and Rust for edge applications and inference pipelines.

The ideal candidate should have practical experience taking computer vision solutions from experimentation to reliable deployment, including integration with backend APIs, cloud services, and dashboards.

Key Responsibilities

Develop and fine-tune models for image classification, object detection, segmentation, and tracking.
Build image enhancement solutions, including denoising, deblurring, low-light enhancement, and super-resolution.
Prepare datasets, manage annotations, apply augmentation, and evaluate model performance across varied conditions.
Train and optimise deep learning models on NVIDIA GPUs.
Develop efficient edge applications in Rust, including image/video handling, inference integration, concurrency, and device communication.
Deploy models on edge processors such as NVIDIA Jetson, Raspberry Pi, and other embedded computing platforms.
Optimise inference using model conversion, quantisation, batching, and hardware acceleration.
Build reusable inference services and persistent workers for processing multiple images or videos.
Develop APIs to expose model predictions and integrate results with cloud applications and dashboards.
Measure and improve accuracy, latency, throughput, memory usage, and deployment reliability.
Maintain documented code and troubleshoot issues across models, applications, and computing infrastructure.

Required Skills

Strong Python programming skills and hands-on computer vision experience.
Working proficiency in Rust, including ownership, error handling, concurrency, and application development.
Experience with PyTorch or TensorFlow, OpenCV, and NumPy.
Practical experience with classification, detection, or segmentation models.
Understanding of model training, transfer learning, loss functions, evaluation metrics, and overfitting.
Experience deploying AI models on GPUs and familiarity with CUDA-enabled environments.
Understanding of edge computing constraints, including limited memory, processing capacity, and power.
Experience with Linux, Git, Docker, and software debugging.
Familiarity with REST APIs, databases, and cloud deployment.

Preferred Skills

Experience with ONNX Runtime, TensorRT, or NVIDIA DeepStream.
Experience integrating model inference into Rust applications.
Knowledge of model quantisation, pruning, and mixed-precision inference.
Experience with NVIDIA Jetson, embedded Linux, or other edge AI hardware.
Familiarity with FFmpeg, video codecs, and camera interfaces.
Experience with image restoration, OCR, vision transformers, or vision-language models.
Knowledge of asynchronous task queues, GPU resource management, and monitoring.
Familiarity with AWS or another cloud platform and dashboard integration.

What We Are Looking For

A developer who can independently investigate problems, write maintainable code, evaluate models objectively, and deliver reliable applications. Strong practical implementation skills and the ability to collaborate across software, AI, and hardware teams are essential.