I'm Eiman

I build AI that actually works. My passion is figuring out how to apply AI to real problems.

Right Now

Head of Solutions @ thirdweb

Building AI agents that actually do stuff

SIWA - Support Agent

Handles customer support. 12 tools. Actually works.

60K+ conversations3/4 CSAT

Billie - Revenue Agent

Handles lead gen and emails. 23 tools. Human-in-the-loop safety.

50% email open rates

Not demos. Production systems. Real customers. Enterprise reliability.

Been at this since 2001

Filed a patent on neural network-based decision systems. Self-organizing maps for product recommendations. This was before neural networks were cool.

Filed: January 2001

Building in Public

Some things I've been experimenting with. AI agents that actually interact with blockchain.

Blockchain Bouncer Agent

AI bouncer that checks NFT ownership before granting access. Combines Eliza agent framework with blockchain verification. Gets your ENS name, verifies VIP Club Pass NFTs, grants access accordingly. Maintains character—multilingual, elite but approachable.

Eliza FrameworkNFT GatingTypeScript

x402 - Agent Commerce

AI shopping agent that actually pays with crypto. LangChain ReAct pattern + x402 payment protocol. Natural language requests turn into real USDC transactions on Base. Chat interface, product search, autonomous purchasing. Not a demo—actual blockchain payments.

LangChainx402 ProtocolUSDC Payments

AI Shopkeeper Game Character

3D medieval marketplace with AI shopkeeper. Turkish bazaar merchant powered by Nebula AI. Walk around as a knight, chat with the NPC, negotiate sword prices. Dynamic pricing based on your token holdings—but the shopkeeper never breaks character to tell you that. Three.js + blockchain + generative AI.

Three.jsNebula AIGaming

Before the AI Hype

Been building production ML systems since before it was cool.

CLO Portfolio Optimizer

First commercially available AI optimizer for CLO managers. Reinforcement learning + linear optimization. Actually shipped.

Featured in Credit Flux for AI innovation

World Economic Forum Recommendation Engine

Built the content rec engine for WEF. NLP + neural nets. Helped them figure out what to write about.

Global reach, data-driven content

Computer Vision at Scale

Face filters and style transfer for a site with 3B+ monthly views. Also built an ML system to parse parking signs. Yes, really.

Billions of interactions, actually in production

How I Got Here

From telecom to hedge funds to fintech. AI consulting. Now blockchain + AI agentic commerce. 30 years of watching tech waves come and go.

The pattern's always the same: people who learn to work WITH the new thing win. Everyone else scrambles to catch up.

Building this community because I kept having the same conversation with students and early-career folks. The frameworks that work are learnable. Figured we should learn together.

Say Hi

I read everything. Questions, ideas, whatever. Reach out.