Nishith Singhai

AI & creative technologist.

less than 1 minute read

🌟 Introduction

  • Problem: building a flexible condition-builder for trading rules.
  • Why I asked ChatGPT for help.
  • Hook: AI + React made it possible faster than I expected.

🧠 The Challenge

  • Needed nested condition blocks (if–else, AND/OR, groups).
  • Hard to code from scratch without clear blueprint.
  • Goal: a UI that even non-devs could use.

🤖 How ChatGPT Helped

  • Prompting ChatGPT with “explain + code sample.”
  • Iterating on JSON structures (RootGroupNode, ChooseBlock, etc.).
  • Debugging logic → refining prompts until output worked.

⚛️ React Implementation

  • Component structure (ConditionBuilder → Group → Block).
  • State management with hooks.
  • JSON export for backend integration.
  • UI design → minimal, intuitive, drag/drop style.

🧪 Testing & Iteration

  • Tried with small test conditions first.
  • Nested conditions tested against mock trading rules.
  • Integration with backend (FastAPI).

💡 Lessons Learned

  • ChatGPT ≠ perfect code → but excellent “pair programmer.”
  • Breaking down into prompts + iterations was key.
  • Future: add drag-and-drop UI, AI-suggested conditions.

🙏 Closing Thoughts

  • Reflection on using AI not just for answers but collaboration.
  • Encouragement: try building your own tools with small AI-guided steps.

Updated:

Comments

You May Also Enjoy

How Modern LLMs Work: A Step-by-Step Guide

2 minute read

From tokens and embeddings to attention heads, probabilities, sampling, and ablations — a practical walkthrough of how large language models process text and...