Website:
kshatralabs.in
Job details:
Kshatra Labs is pioneering the future of autonomous systems, engineering breakthroughs in AI and robotics to build intelligent machines that operate across air, land, sea, and space domains. As a defence robotics company, our primary motive is to build bleeding-edge autonomous robotic systems second to none.
This is a Remote internship role for an Artificial Intelligence Engineer:
Your Responsibilities include:
AI Engineer - Computer Vision & Perception
- Computer Vision: Develop and optimize real-time 3D object detection, segmentation, and tracking algorithms.
- Spatial Perception: Implement camera calibration, stereo vision, and high-accuracy depth estimation pipelines.
- Model Acceleration: Oversee model training and deployment with hardware acceleration using CUDA and TensorRT.
- Frameworks: Utilize PyTorch for large-scale model training and streamlined production deployment.
- Edge Optimization: Deploy and optimize vision models for NVIDIA Jetson and specialized embedded platforms.
- Sensor Fusion: Collaborate on multi-modal fusion algorithms combining camera streams and radar data.
Core Technical Skills:
- Python & PyTorch: Strong coding skills with hands-on project or hackathon experience.
- Computer Vision: Solid grasp of image processing, OpenCV, and standard 2D/3D detection and segmentation models.
- Architecture Theory: Ability to read AI research papers and understand why architectures (e.g., CNNs, ViTs) work, rather than just calling APIs.
- Practical Problem Solving: You know when not to use AI. You can confidently choose classical algorithms or geometry over deep learning when it's more efficient.
- Data Handling: Capable of writing data loaders, cleaning raw datasets, and understanding how data quality dictates model design.
Fundamentals: Comfortable in Linux environments with a basic understanding of how software interacts with hardware.
Development Workflow: Familiar with Git and standard version control practices.
Nice-to-Haves (Bonus):
- Experience with high-speed, low-compute object tracking.
- Exposure to edge computing boards (e.g., NVIDIA Jetson, Raspberry Pi).
- Basic knowledge of model acceleration (e.g., CUDA, TensorRT).
- Background in spatial perception, camera calibration, or sensor fusion (radar/depth).
Click on Apply to know more.