Computer Vision Engineer

We are an early-stage deep-tech startup building real-time AI systems that run on edge hardware. We are looking for a Senior Computer Vision Engineer to be one of our first technical hires — someone who can take a computer vision system from concept to a working product deployed on physical devices, end to end. This is a hands-on, build-everything role. You will not just train models — you will work across the entire pipeline from data to deployment on edge hardware running in real-world environments. If you have shipped a CV product on embedded/edge devices and enjoy working across the full stack, this role is for you. What You Will Do

• Design, train, and fine-tune object detection and tracking models for real-time inference
• Build the complete computer vision pipeline — from video ingestion to detection, tracking, and analytics output
• Optimize models for edge deployment using TensorRT, quantization (INT8), and other acceleration techniques
• Deploy and run models on NVIDIA Jetson devices, handling the full stack from OS setup to production inference
• Build robust video streaming pipelines using RTSP, GStreamer, and NVIDIA DeepStream
• Implement multi-object tracking across multiple camera streams
• Write the application logic and services around the models
• Ensure the system runs reliably unattended in real-world conditions — handling reboots, network drops, and hardware constraints
• Benchmark and continuously improve model accuracy, latency, and throughput
• Set up data annotation pipelines and manage training datasets What We Are Looking For

Must have:

• 4+ years of hands-on computer vision experience
• Strong proficiency in Python and PyTorch
• Deep experience with object detection models (YOLO family, RT-DETR, or similar)
• Proven experience deploying models on NVIDIA Jetson or similar edge hardware — you have personally taken a model from training to running live on a device
• Experience with TensorRT, ONNX, and model optimization for edge inference
• Strong understanding of video processing — RTSP, GStreamer, OpenCV
• Multi-object tracking experience (ByteTrack, DeepSORT, or similar)
• Comfortable writing production-grade code, not just research notebooks

Strongly preferred:

• Experience with NVIDIA DeepStream
• Familiarity with Linux systems programming (Ubuntu/JetPack)
• C++ knowledge for performance-critical components
• Experience building complete products end to end, not just isolated models
• Background in robotics, autonomous systems, drones, IoT, or embedded AI

Bonus:

• Prior experience at an early-stage startup
• Experience setting up MLOps and OTA update pipelines for edge devices
• Open-source contributions or public CV projects Who You Are
• A builder who takes initiative and gets things working, not just theoretically correct
• Comfortable with ambiguity and wearing multiple hats in an early-stage environment
• Able to move fast, prototype quickly, and iterate based on real-world results
• Excited to work close to hardware and see your models run in the physical world What We Offer
• Competitive salary
• The opportunity to build a product from the ground up and shape its technical foundation
• A fast-moving, high-autonomy environment
• Direct impact on a product that runs in the real world

Location: Hyderabad, Telangana, India

Type: Full-time

• Experience: 4+ years