Roshan Patel

I design and deliver AI-powered automation solutions for enterprise clients, combining generative AI, cloud services, and process automation to drive measurable operational improvements. 🔹 AI & Generative Technologies Custom GPT development for automated data processing Azure OpenAI (GPT-4o) integration for context-aware recommendations Azure Document Intelligence with Custom Classifier for document classification and page splitting Confidence-scored field extraction with category-specific schemas and visual validation Prompt engineering strategies for complex data transformation Azure AI Search with semantic search, indexing 5,000+ records Conversational AI workflows for guided user interactions 🔹 Automation & Integration Power Automate Cloud (HTTP triggers, scheduled flows, error handling) Azure Logic Apps for multi-step automation pipelines Multi-layer Power Automate flows connecting AI services, SharePoint, and front-end apps REST API integrations with ServiceNow and external systems JSON parsing, For Each loops, and data synchronization Automated indexing and data pipeline orchestration 🔹 Data Analytics & Visualization Power BI dashboards with real-time API data feeds Live stock analytics tracking 30 Dow Jones companies Interactive visualizations with conditional formatting Power Query for data transformation and cleaning DAX measures for dynamic analysis 🔹 Impact Eliminated manual data transcription across multiple business units through end-to-end AI receipt processing Reduced manual processing time by 40% through AI-powered automation Indexed 5,000+ records for intelligent search and discovery Built real-time dashboards with automatic refresh schedules Streamlined repetitive business processes through AI automation

AI & Robotics Integration Intern focused on developing computer vision systems and automation solutions for industrial spray equipment. 🔹 Computer Vision & AI Developed and optimized real-time computer vision pipelines using YOLOv8 and OpenCV Improved object detection accuracy by 60% during live testing scenarios Automated data workflows with Python for model training and video inference 🔹 Robotics & Hardware Integration Programmed and tested GPIO-controlled robotic components on Raspberry Pi Enabled reliable spray activation with 95% system uptime Configured hardware-software communication for real-time control systems 🔹 Testing & Collaboration Documented performance metrics and collaborated with hardware engineers Improved automation reliability by 15% through systematic testing and debugging Streamlined troubleshooting procedures for integrated systems 🔹 Impact Reduced model iteration time by 30% through automated workflows Achieved 60% improvement in object detection accuracy Maintained 95% system uptime for robotic components