📌Integration
Backend Task – Broker
Layer (One-Way
Sync)
Welcome to the second round of the backend engineering assessment for
Journalyst – our trade journaling platform.
This task is designed to test your ability to design clean architecture, write
modular code, and think through real-world backend systems.
⚠️ IMPORTANT:
Do not use AI tools (like ChatGPT/Copilot) to generate code or make
design decisions.
You are encouraged to refer to official documentation (e.g., Zerodha
Kite Connect API, Alpaca Docs) to understand how real broker APIs work.
Implement it with Typescript, strictly do not use Javascript.
🚀 Goal
Design and implement the backend module that syncs trades from third-party
brokers (eg., Zerodha, MetaTrader etc….). This is a core part of Journalyst's
Automated Broker Sync feature.
You’ll build:
A broker adapter (for one broker)
A data normalizer
Sync logic with token handling
Clear architecture for scaling to other brokers
🧩 What You’ll Build
1. Broker Adapter System
Implement one adapter (e.g., for Zerodha, MetaTrader etc….)
📌 Backend Task – Broker Integration Layer (One-Way Sync) 1
Define a reusable interface
Simulate fetching trade data
NOTE: If you have the ability to generate API keys and tokens, please do
so. If not, simply implement the code as though the API keys and tokens are
already available. The code may not be executable without valid keys, but
we will test it using our own API keys.
2. Trade Data Normalization
You must design the normalization logic yourself.
Real brokers return data in different formats — your task is to study the
structure and convert it into your own consistent format.
Choose relevant fields like symbol, quantity, prices, timestamps, etc.
Do not copy any example schema — use your own skills to build your own
Trade object.
3. Token Management
Simulate per-user token storage (in-memory is okay)
Implement basic token expiry & refresh flow
4. Sync Execution Logic
Create a function: syncTrades(userId, brokerName)
Fetch token → Refresh if needed → Fetch trades → Normalize → Return
trades
Add minimal error handling (invalid token, no trades, etc.)
5. Architecture Design
Structure code into adapters, services, utils, etc.
Make it easy to extend and test
📄 What to Submit
Source code in a public/private GitHub repo
A short README explaining:
📌 Backend Task – Broker Integration Layer (One-Way Sync) 2
Your design decisions
How to add a new broker
Assumptions or simplifications made
Architecture diagram (Markdown, draw.io, or Excalidraw)
(Optional) Sample Postman collection or curl example
(Optional) Walkthrough video link
🧪 Standard Practices We Value
Complete type safety with Typescript
Clean folder structure and modular code
Clear, meaningful function and variable names
Error handling and edge case awareness
Extensibility – your system should be easy to grow
(Optional) Basic testing (even 1–2 tests is appreciated)
💡 Bonus (Not Mandatory)
Provide unit tests for the normalizer or sync logic
Include a docker-compose file for local setup
Add caching or retry logic for failed syncs
📬 Submission
Source code in a public/private GitHub repo
A short README explaining:
Your design decisions
How to add a new broker
Assumptions or simplifications made
Architecture diagram (Markdown, draw.io, or Excalidraw)
(Optional) Sample Postman collection or curl example
(Optional) Walkthrough video link
📌 Backend Task – Broker Integration Layer (One-Way Sync) 3
If you have questions during the task, feel free to reach out — but we expect
you to make key architectural decisions yourself.
📌 Backend Task – Broker Integration Layer (One-Way Sync) 4
