AI Engineer for Crypto SDKs & Agent Ecosystem

CoinMarketCap; Singapore; On-site; SGD 100,000 - 180,000; Full time

Job summary

CoinMarketCap is seeking an experienced platform engineer to design and ship official skills across a 14-domain taxonomy and define SDK ergonomics and revenue-sharing contracts for third-party developers. You will own the MCP server and CLI, making skill authoring as seamless as writing a Vercel function.

You will partner with teams to land CMC skills inside live agent products, implement cross-agent protocol adapters, and feed integration learnings back into the SDK.

Qualifications

• Crypto-native fluency: can explain DEX AMMs, perp funding, bridge risk, on-chain identity.
• Experience contributing to SDK/CLI/API adopted by developers; strong opinions on docs.

Responsibilities

• Design and ship official skills across a 14-domain taxonomy (Leverage, Assets, Portfolio, Pulse, Plays, Risk, etc.).
• Define publishing standards, SDK ergonomics, and the skill-signing/revenue-share contracts for third-party developers.
• Own the MCP server and CLI to simplify skill authoring like writing a Vercel function.
• Partner with teams to land CMC skills inside live agent products; drive cross-agent protocol adapters.
• Co-design and run CMC for Agent hackathons; write technical content and represent at Web3 × AI events.
• Improve LLM grounding for crypto context; build evaluation harnesses for skill accuracy and honesty.

Skills

SDK design

CLI tooling

Agent integrations

LLM grounding

Web3 / AI events

Developer relations

Documentation writing

Solidity basics

Tools

MCP server

CLI

LangGraph

OpenClaw

Job description

CoinMarketCap is seeking an experienced platform engineer to design and ship official skills across a 14-domain taxonomy and define SDK ergonomics and revenue-sharing contracts for third-party developers. You will own the MCP server and CLI, making skill authoring as seamless as writing a Vercel function.

You will partner with teams to land CMC skills inside live agent products, implement cross-agent protocol adapters, and feed integration learnings back into the SDK.