• 🎓 Computer Science grad, currently CTO @ Zhama
• 🌐 Frontend developer at heart who loves Web3 — I ship full products, frontend to backend
• 🤖 Focused on agentic AI: giving LLMs real tools to act on, not just chat
• 📈 Deep into DeFi & prediction-market infrastructure — trading APIs, market makers, AI-driven decisions
• 🔐 Also tinker with networking tools for open, censorship-resistant internet access
• 🌱 Currently exploring the Model Context Protocol (MCP) and agent-run trading systems
• 📫 Reach me at sajadkiyani.ir

I care less about chatbots and more about giving LLMs hands — real tools they can call to observe and act on the world.

• MCP servers for finance — wallex-mcp-server exposes an exchange's market data, wallet balances, and order placement as Model Context Protocol tools, so any MCP-compatible agent can research and trade autonomously instead of me hand-wiring an API integration per use case.
• AI-driven trading decisions — polymarket-ai-decision uses an LLM to reason over prediction-market data and inform trading decisions on Polymarket.
• LLM-powered Telegram automation — telegram-channel-summary is a bot pair: one replies to channel posts with colloquial-Farsi summaries via OpenRouter/Gemini (deny-by-default allowlisted), the other merges an entire Telegram folder of channels into a single daily digest, so I read a folder's worth of news in one message instead of forty.
• Practically that means I spend most of my AI time on prompt-to-tool-call reliability, allowlisting/guardrails for autonomous agents, and structuring LLM output so it's safe to act on unattended.

Frontend background, but most of what I ship these days lives at the trading/market layer:

• Exchange & DEX trading infra — trading-api and lighter-trading-api are high-performance trading APIs against DEX venues, built for latency-sensitive order flow.
• Automated market making — polymarket-marketmaker runs automated market making on Polymarket's prediction markets.
• CEX ↔ agent bridge — wallex-mcp-server (above) is where this overlaps with the AI work: turning a centralized exchange into something an autonomous agent can operate directly.
• Earlier explorations: smart-money-algo (Pine Script) and technical-analysis notebooks, from before I moved to shipping trading infra directly.

🔌 wallex-mcp-server MCP server exposing Wallex exchange market data, balances, and order placement as tools for LLM agents.	🧠 polymarket-ai-decision LLM-driven reasoning over prediction-market data to inform Polymarket trading decisions.
💬 telegram-channel-summary Allowlisted Telegram bot that summarizes channel posts and daily-digests whole folders via OpenRouter/Gemini.	📊 trading-api High-performance DEX trading API built for latency-sensitive order flow.
⚡ polymarket-marketmaker Automated market maker running live on Polymarket's prediction markets.	🏎️ lighter-trading-api High-performance trading API integration for the Lighter DEX.