Job
Description
As a Vision AI Engineer at HCLTech, your role will involve designing, building, and deploying vision-based AI solutions for real-world applications. You will work closely with AI Architects and data engineers to deliver high-performance, production-grade vision systems. Key responsibilities include: - Model Development & Training: - Implement and fine-tune state-of-the-art computer vision models for object detection, classification, segmentation, OCR, pose estimation, and video analytics. - Apply transfer learning, self-supervised learning, and multimodal fusion to accelerate development. - Experiment with generative vision models (GANs, Diffusion Models, ControlNet) for synthetic data augmentation and creative tasks. - Vision System Engineering: - Develop and optimize vision pipelines from raw data preprocessing, training, inference to deployment. - Build real-time vision systems for video streams, edge devices, and cloud platforms. - Implement OCR and document AI for text extraction, document classification, and layout understanding. - Integrate vision models into enterprise applications via REST/gRPC APIs or microservices. - Optimization & Deployment: - Optimize models for low-latency, high-throughput inference using ONNX, TensorRT, OpenVINO, CoreML. - Deploy models on cloud (AWS/GCP/Azure) and edge platforms (NVIDIA Jetson, Coral, iOS/Android). - Benchmark models for accuracy vs performance trade-offs across hardware accelerators. - Data & Experimentation: - Work with large-scale datasets (structured/unstructured, multimodal). - Implement data augmentation, annotation pipelines, and synthetic data generation. - Conduct rigorous experimentation and maintain reproducible ML workflows. Qualifications Required: - Programming: Expert in Python; strong experience with C++ for performance-critical components. - Deep Learning Frameworks: PyTorch, TensorFlow, Keras. - Computer Vision Expertise: - Detection & Segmentation: YOLO (v5v8), Faster/Mask R-CNN, RetinaNet, Detectron2, MMDetection, Segment Anything. - Vision Transformers: ViT, Swin, DeiT, ConvNeXt, BEiT. - OCR & Document AI: Tesseract, PaddleOCR, TrOCR, LayoutLM/Donut. - Video Understanding: SlowFast, TimeSformer, action recognition models. - 3D Vision: PointNet, PointNet++, NeRF, depth estimation. - Generative AI for Vision: Stable Diffusion, StyleGAN, DreamBooth, ControlNet. - MLOps Tools: MLflow, Weights & Biases, DVC, Kubeflow. - Optimization Tools: ONNX Runtime, TensorRT, OpenVINO, CoreML, quantization/pruning frameworks. - Deployment: Docker, Kubernetes, Flask/FastAPI, Triton Inference Server. - Data Tools: OpenCV, Albumentations, Label Studio, FiftyOne. Please share your resume on paridhnya_dhawankar@hcltech.com with the following details: - Overall Experience: - Relevant Experience with Vision AI: - Notice Period: - Current and Expected CTC: - Current and Preferred Location. As a Vision AI Engineer at HCLTech, your role will involve designing, building, and deploying vision-based AI solutions for real-world applications. You will work closely with AI Architects and data engineers to deliver high-performance, production-grade vision systems. Key responsibilities include: - Model Development & Training: - Implement and fine-tune state-of-the-art computer vision models for object detection, classification, segmentation, OCR, pose estimation, and video analytics. - Apply transfer learning, self-supervised learning, and multimodal fusion to accelerate development. - Experiment with generative vision models (GANs, Diffusion Models, ControlNet) for synthetic data augmentation and creative tasks. - Vision System Engineering: - Develop and optimize vision pipelines from raw data preprocessing, training, inference to deployment. - Build real-time vision systems for video streams, edge devices, and cloud platforms. - Implement OCR and document AI for text extraction, document classification, and layout understanding. - Integrate vision models into enterprise applications via REST/gRPC APIs or microservices. - Optimization & Deployment: - Optimize models for low-latency, high-throughput inference using ONNX, TensorRT, OpenVINO, CoreML. - Deploy models on cloud (AWS/GCP/Azure) and edge platforms (NVIDIA Jetson, Coral, iOS/Android). - Benchmark models for accuracy vs performance trade-offs across hardware accelerators. - Data & Experimentation: - Work with large-scale datasets (structured/unstructured, multimodal). - Implement data augmentation, annotation pipelines, and synthetic data generation. - Conduct rigorous experimentation and maintain reproducible ML workflows. Qualifications Required: - Programming: Expert in Python; strong experience with C++ for performance-critical components. - Deep Learning Frameworks: PyTorch, TensorFlow, Keras. - Computer Vision Expertise: - Detection & Segmentation: YOLO (v5v8), Faster/Mask R-CNN, RetinaNet, Detectron2, MMDetection, Segment Anything.
