Arnab Roy

With a Master’s in Artificial Intelligence from IIT Jodhpur and hands-on experience across Generative AI, Computer Vision, Data Analysis, and Data Visualization, my work focuses on building intelligent systems that combine strong AI foundations with real-world business impact. Currently, I am working as a Computer Vision Engineer at Centific Global Solutions, deployed on a Microsoft client project, where I contribute to the development of AI-driven data generation, analysis, and validation workflows. My current work involves designing and improving LLM-based data generation pipelines, applying advanced prompt engineering techniques, analyzing generated data quality, creating use-case-specific validation metrics, and supporting dashboard-driven visibility for stakeholders. I also work on evaluating how well LLMs adhere to structured prompts and business rules, helping improve the reliability and consistency of generated enterprise datasets. My recent experience includes working on Conversational AI systems, including intent detection, slot filling, dataset transformation, synthetic multi-turn conversation generation, LoRA fine-tuning, model evaluation, and error analysis. I have worked with Large Language Models such as Qwen and other open-source models, focusing on practical AI workflows that connect model performance, data quality, and business requirements. Previously, as a Deep Learning Engineer at RadiusAI, I worked on computer vision and visual recognition technologies aimed at improving customer engagement solutions. Before that, as a Product Manager at Futurense Technologies, I gained exposure to MLOps, product development, and the coordination of intelligent systems in data-driven environments. My core interests lie in Generative AI, LLMs, Computer Vision, MLOps, Prompt Engineering, Data Quality, and AI-driven business insights. I am passionate about solving complex problems through applied AI and building reliable, scalable systems that transform data into meaningful outcomes.