CUDA Jobs in Bangalore (Bengaluru)

About the role:

Looking for an engineer to apply Deep Learning algorithms to implement and improve perception algorithms related to Autonomous vehicles. The position requires you to work on the full life-cycle of Deep learning development including data collection, feature engineering, model training, and testing. One will have the opportunity to implement a state-of-the-art Deep learning algorithm and apply it to real end-to-end production. You will be working with the team and team lead on challenging Deep Learning projects to deliver product quality improvements.

Responsibilities:

• Build novel architectures for classifying, detecting, and tracking objects.
• Develop efficient Deep Learning architectures that can run in real-time on NVIDIA devices.

• Optimize the stack for deployment on embedded devices
• Work on Data pipeline – Data Acquisition, pre-processing, and analysis.
•  

Skillsets:

• Languages: C++, Python.
• Frameworks: CUDA, TensorRT, Pytorch, Tensorflow, ONNX.
• Good understanding of Linux and Version Control (Git, GitHub, GitLab).

• Experienced with OpenCV, Deep Learning to solve image domain problems.

• Strong understanding of ROS.
• Skilled with software design, development, and bug-fixing.
• Coordinate with team members for the development and maintenance of the package.

• Strong mathematical skills and understanding of probabilistic techniques.

• Experience handling large data sets efficiently.
• Experience with deploying Deep Learning models for real-time applications on Nvidia platforms like Drive AGX Pegasus, Jetson AGX Xavier, etc.

Add On Skills:

• Frameworks: Pytorch Lighting
• Experience with autonomous robots
• OpenCV projects, Deep Learning projects
• Experience with 3D data and representations (point clouds, meshes, etc.)
• Experience with a wide variety of Deep learning Models (e.g: LSTM, RNN, CNN, GAN, etc.)

Sizzle is an exciting new startup in the world of gaming. At Sizzle, we’re building AI to automatically create highlights of gaming streamers and esports tournaments. 

You will be responsible for:

• Managing all DevOps and infrastructure for Sizzle
• We have both cloud and on-premise servers
• Work closely with all AI and backend engineers on processing requirements and managing both development and production requirements
• Optimize the pipeline to ensure ultra fast processing
• Work closely with management team on infrastructure upgrades

You should have the following qualities:

• 3+ years of experience in DevOps, and CI/CD
• Strong background in Linux system administration
• Deep expertise with AI/ML pipeline processing, especially with GPU processing. This doesn’t need to include model training, data gathering, etc. We’re looking more for experience on model deployment, and inferencing tasks at scale
• Deep expertise in Python including:
• Multiprocessing / multithreaded applications
• Performance profiling including memory, CPU, GPU profiling
• Comfortable with OOP and the nuances of Python classes and other OOP structures
• Error handling and building robust scripts that will be expected to run for weeks to months at a time
• Deploying to production servers and monitoring and maintaining the scripts
• DB integration including pymongo and sqlalchemy (we have MongoDB and PostgreSQL databases on our backend)
• Expertise in Docker-based virtualization including:
• Creating & maintaining custom Docker images
• Deployment of Docker images on cloud and on-premise services
• Monitoring of production Docker images with robust error handling
• Expertise in AWS infrastructure, networking, availability
• Experience with running Nvidia GPU / CUDA-based tasks

• Experience with image processing in python (e.g. openCV, Pillow, etc)
• Experience with PostgreSQL and MongoDB (Or SQL familiarity)
• Excited about working in a fast-changing startup environment
• Willingness to learn rapidly on the job, try different things, and deliver results
• Bachelors or Masters degree in computer science or related field
• Ideally a gamer or someone interested in watching gaming content online