I don't just write code — I build robust, scalable
architectures and orchestrate AI to ship 10x faster.
Combining deep engineering with modern agent
workflows.
I'm Dinudsi, a Full Stack Architect based in Udaipur, India. Since 2009, I've been building enterprise-grade backend systems, secure APIs, and scalable architectures. Today, I augment that deep technical foundation with AI-native workflows to help startups and companies get their MVPs done in days rather than months, with an 80% cost reduction.
As an engineer first, I rely on tools like Spring Boot, Node.js, and PostgreSQL to build the core, while leveraging Claude Code, Codex, and Copilot CLI to accelerate execution.
Dinudsi combines 13+ years of traditional software architecture with modern multi-agent orchestration. He builds open-source plugins, custom MCP servers, and internal dev tools like BrainSync to push the boundaries of what a single developer can achieve. His focus remains on security-first code review, clean architecture, and shipping production software rapidly.
Dinudsi's toolkit is built around AI coding agents, custom MCP servers, agent orchestration, and security-first code review — designed for shipping production MVPs at unprecedented speed.
AI agent tools, custom MCP servers, open-source plugins, and production systems — built by Dinudsi to push the boundaries of AI-native development.
A small transformer-based language model built from scratch using PyTorch. Trained to generate structured tool-calling responses via next-token prediction — demonstrating core ML engineering skills.
Cross-agent memory and context management system. Enables AI coding agents to retain knowledge, share context across sessions, and build persistent memory using vector embeddings.
Suite of Model Context Protocol servers that extend AI agent capabilities — from file operations and code analysis to custom tool integrations and workflow automation.
Role-based agent management system for development workflows. Coordinates specialized agents for code generation, security review, testing, and deployment — all working in concert.
Real-time AI system for vehicle plate recognition, violence detection, and threat analysis. Integrated vision-language models (VLM) for contextual live video understanding.
Production-ready OAuth2 JPA-based authorization server with client registration, token lifecycle management, user approval flows, and role-based authorities.
VS Code extension that automatically commits your changes every 5 minutes using GitHub Copilot to generate intelligent commit messages. AI-powered developer productivity tool.
How Dinudsi leverages AI agents, custom tooling, and security-first practices to deliver production software at 10x speed.
Runs role-based agents for code generation, review, testing, and deployment — coordinated through custom orchestrators.
Builds and maintains MCP servers that extend AI agent capabilities with custom tools, integrations, and workflows.
Designed persistent memory and context management tools (BrainSync) enabling agents to retain knowledge across sessions.
Every PR reviewed with AI-assisted vulnerability detection, secret scanning, and threat modeling before merge.
I don't just write code — I orchestrate AI systems that write, review, and ship code.
I am available to take on new MVPs and love working
on challenging, out-of-the-box ideas.
Let's build something amazing together at
unprecedented speed and reduced cost.
Dinudsi is a Full Stack Architect and AI-Native Developer based in Udaipur, Rajasthan, India. With 13+ years of software engineering experience since 2009, he combines deep expertise in traditional software architecture with modern AI-augmented workflows.
His core engineering skills include Spring Boot, Node.js, PostgreSQL, Docker, AWS, and secure OAuth2 implementation. He augments this foundation by orchestrating AI coding agents like Claude Code, Codex, and Copilot CLI to ship production software rapidly. He also builds custom MCP (Model Context Protocol) servers and agent memory systems to extend AI capabilities.
Notable projects include BrainSync (a cross-agent memory system), BrainByte (a transformer-based language model built from scratch), a custom MCP server ecosystem, an AI-Enabled CCTV Security System, a production OAuth2 Authorization Server, and Commitify (an open-source VS Code extension for AI auto-commits).
His philosophy bridges robust system design with AI-accelerated execution, focusing heavily on security-first code reviews and scalable architecture. Contact: dpanchal.fullstack@gmail.com | GitHub: github.com/dinudsi