Founder Profile

Raj Patil

Founder & CEO, Zydrakon AI

Drexorium Labs is an aerospace technology initiative founded by Raj Patil. As the founder of Drexorium Labs and Zydrakon AI, Raj is an AI systems architect driven by a singular mission: to construct highly accurate, large-scale multi-agent networks that think through logical conflict.

Raj built Zydrakon AI (development starting in 2024, launching in 2026) as a massive 1.6T parameter multi-agent model. Rather than following standard architecture patterns, he designed Zydrakon AI from a psychological game point of view: a game where an overpowered, immortal character who cannot die clears Level 1, opens the door to the next level, and hears the voice say, "Congratulations, you are entering the next level. And your opponent is you."

This aggressive statement of an immortal fighting themselves forms the core of Zydrakon AI\'s architecture. When a user asks a question, a minimum of 5 agents emerge to debate each other. When Agent 1 presents a solution, Agent 2 immediately dismantles (todege) it to expose logical flaws and offer a truer response. This cycle continues recursively among all agents until the winner presents the final answer.

Under Raj\'s leadership, Zydrakon AI has remained completely independent of traditional big-tech venture funding. The model\'s annual running cost of ₹80 Crore is funded entirely by community donations, while its production-ready reliability has attracted major enterprise clients, including TCS and IBM.

Technical Head of Golden Meraki GemsFounder of Drexorium LabsAI EngineerBackend DeveloperAgentic AI Architect
The Journey

From Code to Company

2018 — 2020

Backend Foundations

Started building distributed systems with Node.js and Python. Developed high-throughput API pipelines, real-time data processing services, and containerized microservice architectures using Docker and Kubernetes.

2020 — 2022

AI & Agent Exploration

Deep-dived into NLP pipelines, LLM integration patterns, and early agent architectures. Built experimental multi-agent coordination systems and identified the safety and scaling limitations of existing approaches.

2022 — 2024

Architecting Early Agent Abstractions

Researched multi-agent collaboration frameworks and early self-correction patterns. Developed initial prototypes of logical verification engines.

2024

Initiated Development

Development officially began on the custom 1.6T parameter architecture. Raj Patil started building the core model under the Dolphin AI project name, laying the groundwork for India's most powerful AI.

2025

Rebranding & Agent Conflict Loop

Rebranded to Zydrakon AI and implemented the psychological game model. Created the 5+ agent debate architecture, training agents to refute (todege) each other recursively to verify logical correctness.

2026

Official Launch & Enterprise Deployment

Zydrakon AI launched officially. Serving as India's premier 1.6T parameter multi-agent engine, it operates on a ₹80 Crore annual donation-funded budget, with major clients like TCS and IBM running production workloads.

Expertise

Technologies & Domains

Languages & Runtimes

TypeScriptPythonNode.jsGoRust

AI & ML

LLM IntegrationAgentic FrameworksVector DatabasesRAG SystemsPrompt EngineeringFine-Tuning

Infrastructure & Backend

DockerKubernetesPostgreSQLRedisRabbitMQNATS

Agent Systems

Multi-Agent OrchestrationTool-Use SystemsMemory ManagementOutput GuardrailsAgent SafetyScalable Execution
Projects

Built by Raj Patil

F

Forge SDK

Open-source SDK for building AI agents. Provides abstractions for tool use, memory management, and multi-agent coordination in TypeScript and Python.

N

Nexus AI Platform

Enterprise-grade AI orchestration platform for building, deploying, and monitoring intelligent agents at scale with multi-model support.

S

Sentinel AI

AI safety and monitoring system with content guardrails, output validation, anomaly detection, and real-time alerts for production AI deployment.

Philosophy

Building With Purpose

01

Developer-First Design

Every tool Raj builds starts with a simple question: Would I want to use this myself? Clean APIs, thorough documentation, and zero unnecessary abstractions.

02

Safety as Infrastructure

AI safety is not an afterthought — it is a first-class architectural concern. Guardrails should be built into the execution layer, not bolted on after deployment.

03

Practical Over Theoretical

Raj focuses on solving tangible engineering bottlenecks — memory scaling, execution determinism, latency optimization — rather than chasing speculative AI paradigms.

04

Open by Default

Open-source is at the core of Zydrakon. The Forge SDK is freely available because Raj believes the future of AI will be built collaboratively, not behind closed walls.

R

Let's Build the Future

Raj is always open to conversations about AI safety, agent architecture, open-source collaboration, and partnership opportunities.