GitHubAgentic Systems
Autonomous agents that plan, collaborate, use tools, remember context, and execute complex tasks.
Independent Security Researcher & Developer
Building safer autonomous AI systems that people can trust.
I'm an independent security researcher and developer building multi-agent systems, AI security tools, and shipped consumer apps focused on reliability, interpretability, and real-world impact.
AI AgentsAI SafetySecurityResearch
AI
What IFocus On
Practical systems for safer, more inspectable autonomous intelligence.
AI Security
Red-team workflows, prevention agents, trust layers, and human-in-the-loop verification.
Research Direction
Interpretability, multi-agent coordination, safety evaluation, and transparent decision traces.
CurrentResearch
A focused prototype for agent governance, shared memory, and auditable autonomous software delivery.
Orvix - Research Prototype
One product request becomes a self-organizing AI agency.
Orvix is my attempt to build a temporary software company made of agents. The user gives a high-level product mission, and the system decides which agents are needed, what each one owns, what can happen in parallel, what must wait, how review should work, and when the final output is ready.
Live coordination layer
MasterMind Agent
HackathonWinner Builds
Recognized projects built end-to-end under competition deadlines. Click any project for the full story.
Building& Shipping
Alongside security research, I design and ship complete consumer products end-to-end.
Meizhi
Smart Recipes. Simple Cooking.
Meizhi takes "what should I eat?" off your plate. Tell it your preferences, what's in your fridge, and how confident you are in the kitchen, and it builds a full day of tailored meals - breakfast to snacks - with complete recipes, nutrition estimates, and a visual of each dish.
Not a medical or nutrition service - meal suggestions and nutrition values are AI-generated estimates.
Nyvo
A new version of you, every morning.
Nyvo turns your selfies and inspiration into a personalized AI wallpaper. Choose Custom mode, or let Nyvo's Auto mode decide - an autonomous backend runs every day at 2am, generates a fresh wallpaper, and a push notification wakes the device to download and apply it before you wake up.
Research Direction
Autonomous, but never a black box.
As an independent security researcher and developer, I build agents that can act independently while staying inspectable, controllable, and aligned with human intent. Across security investigation, blockchain intelligence, developer tooling, and structured reasoning, the throughline is the same: let AI act on its own without losing the ability to verify what it did and why.
Current direction
Human-verifiable multi-agent systems with traceable decisions, safety checks, and audit layers for real-world deployment.
Let's build something
Curiosity builds the future.
Curiosity builds the future.
Systems make it real.
Open to research collaborations, fellowship opportunities, and AI safety conversations.
