Technical AI Product Manager: CoinMarketCap

Role Overview
This is not a traditional Product Manager role where you write PRDs and manage Jira tickets. This is a Prototyper / AI Engineer-PM Hybrid.

The job description explicitly states: "You are responsible for producing strong prototypes... identifying user pain points, prototyping in Python." You are being hired to be the "Zero to One" builder. You must personally write the Python code to build a working internal demo of an AI feature (e.g., "Explain why Bitcoin is down today"). Once your prototype works and adds value, you hand it off to the core Engineering team to make it scalable and production-ready.

Key Responsibilities
The Builder: You will spend a significant amount of time in VS Code or Jupyter Notebooks. You need to verify if an idea is possible by building it yourself.

Prompt & Context Engineering: You are responsible for the system prompts. You must figure out how to feed CoinMarketCap’s massive data stream (prices, news, on-chain data) into an LLM context window without blowing the budget or confusing the model.

Evaluation Loops (The Hard Part): You must define "Golden Sets" (perfect answers) to test the AI. If the AI says "Solana is a scam," you need a regression test to catch that before it goes live.

Handoff: You don't manage the servers, but you define the logic. You give Engineering the "blueprint" (your Python prototype) rather than just a text document.

Strategic Context: Why this role exists
CoinMarketCap (CMC) is the most trafficked website in crypto, but it is currently just a wall of numbers.

The Problem: Users see a token is up 50%, but they don't know why.

The Solution: They want an AI that acts as a customized financial analyst for every user, synthesizing news, on-chain data, and price action into a coherent narrative.

The Parent Company: CMC was acquired by Binance. Expect the Binance culture: extremely fast-paced, direct feedback, and a demand for high execution speed ("High-agency execution").

Candidate Profile: The "Hacker PM"
You are the ideal candidate if:

You are a PM who actually codes. If you cannot write a Python script to call the OpenAI API and parse a JSON response, do not apply.

You understand RAG (Retrieval-Augmented Generation). You know that you can't just ask GPT-4 about a crypto price; you have to fetch the price from a database and inject it into the prompt.

You are a Crypto Native. You understand what "on-chain volume" or "FDV vs. Market Cap" means.

Locations & Remote Work
While it says "Global," the listing highlights Dubai, Hong Kong, and Singapore. This usually implies they prefer candidates who can work in UTC+4 to UTC+8 time zones, or who are willing to work late nights if based in the US/Europe.