onsite
Computer Vision Engineer PyTorch/TensorRT - Flatgigs
Software Engineer
Computer Vision Engineer focused on training and fine‑tuning detection, classification, and segmentation models (YOLO, ResNet, U‑Net) using PyTorch, optimizing them with NVIDIA TensorRT, and deploying production‑grade Python code in Docker containers for scalable, high‑performance AI solutions.
About the role
We are seeking a Computer Vision Engineer with strong software and AI fundamentals to build and deploy high-performance AI models. You will handle the full pipeline—from training detection and segmentation models to optimizing them for production using NVIDIA TensorRT and Docker.
Core Responsibilities
- Model Training: Train and fine-tune models for Detection, Classification, and Segmentation (e.g., YOLO, ResNet, U-Net).
- Tracking: Implement Multi-Object Tracking (MOT) algorithms for complex video streams.
- Engineering: Write production-grade Python code with a focus on modularity and scalability.
- Deployment: Containerize applications using Docker for consistent deployment.
Requirements
- 3+ years in CV/Deep Learning.
- Python, PyTorch, OpenCV.
- Strong preference for experience with NVIDIA TensorRT and model optimization (quantization/pruning).
- Solid grasp of software engineering principles (Git, testing, CI/CD).
- Can work on other non-vision AI implementations
Originally posted on Himalayas
Skills
pytorchdockerpython