2026-02-15

project1

project 1: the “order book” aggregator

scenario: a crypto exchange (like the one you worked on at blazpay) is crashing because the frontend is polling the database for prices. the mission: build a real-time websocket ingestion engine that normalizes order book data from binance & coinbase into a unified kafka stream.

• tech: python (asyncio), kafka, docker.
• challenge: handle connection drops and ensure strict ordering of updates.
• dev to prod:
  1. create a python aiohttp script to connect to binance websocket.
  2. push raw json events to a kafka topic raw_orders.
  3. prod requirement: dockerize the producer with a restart: always policy. implement “dead letter queue” logic (if kafka is down, save to a local csv buffer).