Founding AI Engineer

About the Company

We are building AI-powered industrial machines. We believe manufacturing is an art form that has been left behind. Our mission is to give this art the tools it deserves by building machines that can learn, adapt, and improve. We're creating factories that can think. We are a small, ambitious team building from first principles, and we're looking for people who want to shape the future of manufacturing.

About the Role

We are building intelligent machines that interact with the physical world. As a Founding AI Engineer, you will design and deploy the AI models that run at the edge, powering real-time decisions in complex industrial environments.

This is a chance to build computer vision models that power real machines in the real world. You will be part of a mission-driven, engineering-first culture, helping build a hardware startup in India from the ground up.

Responsibilities

• Develop and train deep learning models for object detection, segmentation, tracking, and defect classification.
• Build depth-based inspection systems using stereo vision and structured light.
• Optimize models for high-performance, real-time inference on edge devices.
• Build and manage data pipelines for collection, labeling, augmentation, and retraining.
• Collaborate closely with optics, embedded, and software engineers to deploy models in the field.
• Evaluate and benchmark system performance in real-world industrial environments.

Requirements

• 3-4 years of hands-on experience building and deploying computer vision or AI systems.
• Hands-on experience with the NVIDIA stack (e.g., DeepStream, GStreamer, Triton Inference Server, TensorRT) to build low-latency, high-throughput vision pipelines.
• Proficiency in Python, PyTorch/TensorFlow, and OpenCV.
• Experience with image/video data, augmentation, and preprocessing.
• A good understanding of camera geometry, image noise, and lighting variation.

Good to Have

• Experience with stereo vision, structured light, or depth reconstruction.
• Experience with model quantization, pruning, or edge deployment (e.g., TensorRT).
• Familiarity with MLOps or training data pipelines.
• Solid understanding of 1D/2D metrology or classical computer-vision techniques.
• Experience writing or customizing CUDA kernels to accelerate inference