Job
Description
Job description Job title: Consultant - Data Scientist- Computer Vision Location: Bangalore (Hybrid) Years of Experience: 5+ Preferred - Immediate Joiners Primary Skills: Computer Vision (YOLO, PaddleOCR, Video/Image Analytics, number plate identification, parking spot detection, detecting speed of vehicles from videos) Secondary Skills: Neural Networks, Computer Vision, Deep Learning Key Responsibilities: AI/ML Model Development & Optimization Design, develop, and train deep learning models, including CNNs, RNNs, Transformers, and hybrid models for audio-video data fusion. Optimize models for performance, including reducing inference times and memory footprint for deployment on resource-constrained devices. Data Collection & Preprocessing Collect, preprocess, and augment large-scale datasets, particularly for audio and video, ensuring high-quality input for model training. Implement data fusion techniques to combine multimodal data sources (audio, video) to create richer, more accurate models. Edge Computing & Deployment Deploy AI/ML models onto edge devices such as Raspberry Pi, NVIDIA Jetson, or Intel Movidius, ensuring real-time, efficient operation. Perform hardware-aware optimizations like quantization, pruning, and model compression to enable efficient edge deployment. Real-time AI Applications Develop real-time AI applications using computer vision and deep learning techniques for tasks such as object detection, gesture recognition, and video analysis. Ensure high accuracy and low latency for real-time performance, both in cloud and on-device environments. Collaboration with Cross-Functional Teams Collaborate with software engineers, hardware engineers, and data scientists to integrate AI models into embedded systems, ensuring smooth deployment and functionality on edge devices. Work closely with product teams to translate business requirements into AI solutions, ensuring the models meet end-user needs. Model Evaluation & Monitoring Evaluate the performance of deployed models using standard metrics (accuracy, precision, recall, F1 score) and real-world feedback to improve model performance over time. Continuously monitor and troubleshoot models post-deployment to ensure optimal performance and scalability. Research & Innovation Stay up-to-date with the latest advancements in AI/ML, computer vision, and edge computing, and explore new techniques to improve the efficiency and effectiveness of models. Experiment with novel architectures, including fusion models and multi-task learning, to solve complex problems in multimodal environments.
