Computer Vision Intern

**Skill Requirement**

**Background**:
Praan (Praan, Inc.) is an impact focused deep-tech startup democratizing clean air using breakthrough filterless technology. The company is backed by top tier VCs and CXOs globally and currently operates between the United States and India. Our team puts extreme attention to detail and loves building technology that's aspirational. Praan's team and culture is positioned to empower people to solve large global problems at an accelerated pace.

**Why**:
Everyone worries about the dooms-day in climate change which is expected to occur in the 2050s. However, there's one doom's day which is the reality for millions of people around the world today. Air pollution takes more than 7 Million lives globally every single year. Over 5% of premature children death occur due to air pollution in developing countries. Everyone has relied on governments or experts to solve the problem, but most solutions up until today have either been too expensive or too ineffective. Praan is an attempt at making the future cleaner, healthier, and safer for the generations to come.
- Implement advanced 3D reconstruction pipelines for point clouds, following the latest research, using Python and C++
- Develop and deploy deep learning models, including transformer-based models, to enhance the functionality and performance of our products
- Optimize algorithms for low-latency and high-efficiency performance on embedded platforms and edge devices
- Design, develop, and train machine learning models for automatic clean-up and transformation of 3D point clouds into CAD files
- Write and maintain custom bash scripts to support automation and workflow optimization
- Work closely with mechanical design, manufacturing, embedded system and CFD teams to create integrated physical systems
- Work with deployment teams to field-test computer vision systems in diverse factory environments, refining performance and reliability
- Create custom tools and algorithms for seamless conversion of point cloud data into CAD formats with enhanced usability for design and manufacturing teams
- Implement real-time processing pipelines for point cloud and vision data to enable immediate insights and decision-making
- Prepare high quality documents, manuals, and test reports for project scaling
- Conduct rigorous integration testing to ensure seamless compatibility between hardware, software, and firmware components

**Skill Requirements***:

- BS/MS in Computer Science, Electrical Engineering, or a related engineering field
- Past experience in Computer Vision, Perception, and Mapping in indoor environments
- Strong proficiency in Python, with hands-on experience in training object detection and segmentation models using TensorFlow or PyTorch
- Experience with transformer architectures for tasks such as audio classification and video context extraction
- Past Experience in working with LiDARs, 3D point clouds, segmentation and clustering
- Experience in 3D reconstruction, Pointclouds/Image data to CAD
- Hands-on experience on embedded devices like Nvidia Jetson, Raspberry Pi, etc
- In-depth knowledge and hands-on experience with Linux environments

**Praan is an equal opportunity employer and does not discriminate based on race, religion, caste, gender, disability or any other criteria. We just care about working with great human beings