Sr. Computer Vision Engineer (Pytorch)

Visionify is working on bringing the power of Computer Vision and AI for everyday use-cases. We are looking to hire a strong, motivated and enthusiastic Sr. Computer Vision Engineer to execute our roadmap.

As a Sr. Computer Vision Engineer, you will be working on the state-of-the art challenges in the field of Computer Vision & solving them with novel algorithms and optimizations. Majority of our work is focused on applied Computer Vision - so a good understanding of current state of the art of models (Classification, Object detection, Object Recognition, OCR, LayoutML, GAN etc networks). You will be working on Pytorch as primary language - so prior demonstrated knowledge of Pytorch is must for this position. Any experience with Azure, Azure ML Studio framework etc., would also be preferable.

Candidates are expected to stay current with the latest features and contribute to the open-source Pytorch project with a focus on performance and accuracy improvements. You deeply understand the PyTorch framework and underlying implementations to solve customer challenges., and provide insights into how key issues affect the product. Many of our models get deployed to the edge - so experience in optimizing and pruning models, converting models to NVIDIA TensorRT etc would be preferable.

You must possess excellent Python coding skills - as Python is used throughout our organization to build training and inference pipelines. Candidates should have excellent communication and presentation skills. The ideal candidate will be passionate about artificial intelligence and stay up-to-date with the latest developments in the field.

Responsibilities :

Understanding business objectives and developing Computer Vision based solutions that help achieve it. The software could involve training framework, inference framework, working with different technologies for ML.

You will build models and solutions with Pytorch. You will optimize the Pytorch models for different runtime environments including NVIDIA Jetson TensorRT.

Guide the development team with their works, unblock their questions, help accelerate their deliverables.Developing ML / Computer Vision algorithms that could be used to solve a given problem.

Exploring and visualizing data to gain an understanding of it, then identifying differences in data distribution that could affect performance when deploying the model in the real world

Develop processes for different common operations of the team : data acquisition, model training, prototype development.

Finding open-source datasets for prototype development.

Develop pipelines for data processing, augmentation, training, inference and active retraining.

Training models and tuning their hyperparameters

Analyzing the errors of the model and designing strategies to overcome them

Deploying models to production

Requirements :

Bachelors / Masters degree in Computer Science / Computer Engineering / IT or related fields.

5+ years of experience. Exceptional candidates with less experience are welcome to apply.

Industry experience working in Image & Video Processing (OpenCV, GStreamer, Tensorflow, PyTorch, TensorRT, Model Training / Inference, Video Processing Pipelines, Different GStreamer Convertors etc).

Sound knowledge of various deep learning classification models including ResNet, Inception, VGG etc and object detection models including MobileNetSSD, Yolo, FastRCNN, MaskRCNN etc.

Good knowledge of Pytorch, Torchvision, writing training routines. Ability to update models, add / drop features, visualize how the model is performing etc.

Experience in Colab and Jupyter Notebook

Familiarity with CUDA / GPU

Knowledge of CNN visualization techniques such as CAM, GradCAM etc.

Strong understanding of Computer Vision and Real-time Video Processing techniques.

Strong experience with Python and writing reusable code.

Experience working with OpenCV and Scikit packages.

Experience with the NVIDIA platform (NVIDIA Deepstream, TensorRT).

Experience with Python web framework, e.g. Flask, Django or FastAPI

Experience with different ML platforms : PyTorch, TensorFlow.

Proficiency with AWS SageMaker

Experience with databases (Elasticsearch, SQL, NoSQL, Hive, …)

Experience in a cloud environment for software development and deployment (AWS preferred)

Experience with utilising various GPU-based training infrastructures.

Experience with Docker

Knowledge of DevOps and MLOps best practices for production Machine Learning systems.

Desired Traits :

Thrive in a collaborative environment.

Flexible with changing requirements.

Come up with innovative solutions.

Keen focus on work quality & developing robust code.