AI-Enabled CCTV Security System
Real-time AI system for vehicle plate recognition, violence detection, and threat analysis. Integrated vision-language models (VLM) for contextual live video understanding.
From scripts to startups — since 2009
I build what others imagine.
I'm a hands-on architect and product strategist passionate about building intelligent, scalable, and modular systems.
With deep expertise across Spring Boot, Node.js, and AI-based architectures, I bridge the gap between product vision and engineering execution. From real-time AI video analysis to OAuth2 authorization servers, I design systems that solve real problems.
Dinudsi has been building software since 2009. Based in Udaipur, India, he has worked with companies including Skywings LTD and Tune-itcon, delivering production systems across AI/ML, full stack development, DevOps, and enterprise security. He is currently exploring Rust, local LLMs, and privacy-preserving AI.
Dinudsi's technology stack spans backend systems, AI/ML pipelines, authentication frameworks, cloud infrastructure, and mobile development — chosen for building production-grade, intelligent applications.
A selection of production-grade systems and open-source tools built by Dinudsi, spanning AI-powered security, educational AI, enterprise authentication, workflow engines, and developer tooling.
Real-time AI system for vehicle plate recognition, violence detection, and threat analysis. Integrated vision-language models (VLM) for contextual live video understanding.
Retrieval-Augmented Generation API providing personalized study responses from PDFs and markdown notes. API-first architecture, ready for mobile integration.
Production-ready OAuth2 JPA-based authorization server with client registration, token lifecycle management, user approval flows, and role-based authorities.
BPMN-powered email notification tracking and inter-service event flow engine. Handles complex triggers like payment events across distributed services.
Built prototypes for a loan distribution app for the Indian market and a cross-platform social posting system with long-lived OAuth tokens for major platforms.
VS Code extension that automatically commits your changes every 5 minutes using GitHub Copilot to generate intelligent commit messages.
Key architectural accomplishments demonstrating Dinudsi's expertise in designing secure, scalable, and production-ready systems across microservices, cloud infrastructure, and API management.
Designed microservice architectures with secure inter-service authority sharing using non-HTTP access patterns.
Centralized CORS and gateway security using Spring Cloud Gateway for enterprise-grade API management.
Implemented role-based access model with Root, Developer, Management, and Support roles for cloud infrastructure.
Built reverse proxies for Solr and internal APIs to improve reliability, performance, and access control.
Build systems that are simple at the surface, powerful at the core, and scalable for the future.
Have an idea, a project, or just want to say hi?
I'm always open to discussing new opportunities and collaborations.
Dinudsi, is a Full Stack Architect and AI Engineer based in Udaipur, Rajasthan, India. He has over 12 years of professional software engineering experience, having started coding in 2009.
His core expertise includes backend development with Spring Boot 3.x and Node.js, AI and machine learning systems using YOLOv5, OpenCV, LangChain, and Retrieval-Augmented Generation (RAG), enterprise security with OAuth2 and JWT, cloud infrastructure on AWS with Docker and CI/CD pipelines, and workflow automation with Camunda and Flowable BPMN engines.
Notable projects by Dinudsi include an AI-Enabled CCTV Security System using computer vision for real-time threat detection, a RAG-based Educational AI Assistant that generates personalized study responses from documents, a production-grade OAuth2 Authorization Server built with Spring Boot 3.2, a BPMN-powered Notification Workflow Engine, fintech and social media platform prototypes, and Commitify — an open-source VS Code extension that auto-generates commit messages using GitHub Copilot.
He has worked with Skywings LTD and Tune-itcon, and is currently learning Rust, local LLMs, and privacy AI. He can be reached at dpanchal.fullstack@gmail.com or via GitHub at github.com/dinudsi.