Hyunjun Kim
← All work

Case 08 / 09

Trading & Kimchi-Premium Monitoring Bots

Built with AI: bots that track crypto prices and the Korea premium (kimchi premium) and alert via Telegram

Experiment

2024.12 – 2025.01 (kimchi-premium bots) · 2026.02 – 04 (trading bots) / Solo (AI as the team) / Server / bot

Problem
Watching crypto prices and the kimchi premium by hand is inefficient.
My role
Planning · directing AI to build
Stack
Python · requests · pandas · Flask · Telegram Bot API

Results

Real-time alerts delivered automatically
Telegram
Real-time alerts delivered automatically
Monitors domestic vs overseas price gaps
Kimp tracking
Monitors domestic vs overseas price gaps
₿Price Alert BotExample · dummy data
  • Premium alert · BTC

    Domestic vs overseas gap +3.2% (above 3.0%)

    09:41

  • Price move · ETH

    −4.1% in 1h (hit −4.0%)

    10:05

  • Status

    Bot running (Docker)

    12:00

01

Idea / Problem

Watching crypto prices — and the gap between Korean and overseas exchanges (the "kimchi premium") — by hand is inefficient. I built a bot that detects changes automatically and alerts via Telegram.

02

The Plan & Bar I Set + Directing AI

  • Bar: alert through Telegram (already in use) instead of a separate app/dashboard; if the bot stops, alerts stop, so run it always-on with Docker. I planned the pipeline — price/kimp collection, condition checks, alerts — and directed AI step by step.

03

Results

  • Collects prices and kimp and sends automatic Telegram alerts when conditions are met
  • Iterated across several versions (v2 includes the Docker setup)

04

Where It Got Stuck & How I Resolved It

  • After an early loss from over-trusting trade logic, I adopted a principle: explain the impact and get confirmation before changing anything. I still apply it when directing AI on my other projects.

05

Retrospective

Small in scale, but it taught me both the value of automation and that "auto-trading demands caution." I now use it only for information gathering and alerts, not automated trading.

Note: A personal, experimental project — a supporting example of Python automation and server operations.

Next project →

MemoHub

A web app that gathers scattered notes and Telegram messages, analyzes/summarizes them with AI, and even gives same-day reminders — deployed

MemoHub screenshot
View case
Resume PDF