Rhythm Vohra

Innovative and results-driven AI/ML professional with over six years of combined industrial and research expertise in developing and deploying cutting-edge machine learning models. Successfully designed, optimized, and managed 200+ AI models, demonstrating a strong ability to solve complex problems with a collaborative approach. Recognized for pioneering segmentation models and custom loss functions that push the boundaries of artificial intelligence. A published researcher with two papers at esteemed international conferences (with two more in progress), actively contributing to the AI/ML community through presentations and peer reviews. Passionate about advancing AI-driven solutions through rigorous research, strategic implementation, and cross-functional collaboration.

• Experienced AI/ML Engineer with a strong background in research and industry, specializing in computer vision, deep learning, and large-scale AI solutions. At Picsume Inc., led the development of an AI-driven job recommendation engine using LLMs and RAG, implementing a twin-tower approach with vector embeddings to enhance matching accuracy. Integrated the system into Google Cloud Platform (GCP), ensuring scalability through MLOps best practices.
• As a Graduate Research Assistant at the University of Victoria’s Computer Vision Lab, collaborated with Quirklogic on single-class instance segmentation for line drawing vectorization. Developed a novel segmentation model and loss function that surpassed the state-of-the-art in both processing speed and accuracy. Published findings at VISIGRAPP 2024, receiving high praise from reviewers.
• At ASL Environmental Sciences, researched semantic segmentation for detecting underwater scatterers in echograms, with work presented at CVPR 2023. Also explored deep learning-based echogram annotation methods, contributing to acoustic classification research.
• Previously, at UnitedHealth Group, developed AI-powered automation solutions, NLP-based document processing pipelines, and over 30 REST APIs. Managed a team of four, contributing to large-scale AI projects. Proficient in deep learning, NLP, MLOps, and cloud technologies, with a track record of bridging research and production-ready AI solutions.

Master of Applied Science in Electrical and Computer Engineering, University of Victoria:

Conducted cutting-edge research in computer vision and deep learning, focusing on single-class instance segmentation for line drawing vectorization. Developed a novel segmentation model and loss function, significantly improving processing speed and accuracy over state-of-the-art methods. Research was published at VISIGRAPP 2024 and received high praise from reviewers. Specialized in AI/ML methodologies, including large-scale model deployment, NLP, and advanced deep learning techniques