Chan Inthisone
Back to Projects

Quant Research Platform

QuantMLTrading

My journey building an end-to-end quant research pipeline.

Quant Research Platform

Background

I found Engineers Gate's work on systematic equity strategies and got hooked. I dove down a rabbit hole. I learned quant dev was a steep curve. Web dev, data engineering, DevOps, finance, math—all in one field. I realized quants don't start as experts. They learn as they go. I thought, "Can I do this?" I love a challenge. I organized notes in Obsidian and saved code and formulas in Markdown. That helped me learn concepts quickly.

Pipeline

+--------------------+
|  Data Ingestion    |
+--------------------+
          |
          v
+--------------------+
|  Data Lake (Raw)   |
|  - CSV / Parquet   |
+--------------------+
          |
          v
+--------------------+
| Feature Engineering|
| - Rolling Returns  |
| - Yield Curve Diff |
| - Sentiment Scores |
+--------------------+
          |
          v
+--------------------+
|     Modeling       |
| - Prophet / LSTM   |
| - Regime Clustering|
+--------------------+
          |
          v
+--------------------+
|     Backtesting    |
| - Custom Strategy  |
| - Metrics Output   |
+--------------------+
          |
          v
+--------------------+
|    Signal Output   |
| - Buy / Sell Flags |
+--------------------+
          |
          v
+--------------------+
| Execution (Opt.)   |
| - Alpaca API       |
+--------------------+

Challenges & AI Help

I hit countless errors building this. Nights went by debugging. AI was my lifeline. It guided me through model tweaks and data bugs. When it finally ran, I felt unstoppable.

Results & Next Steps

I backtested 2012–2024 data. The strategy returned 12% CAGR with a Sharpe of 1.3. I plan to revisit each code section using AI to deepen my understanding. Next, I'll add IBKR integration and learn to code strategies in C++. I'm proud I pulled this off. I'm just getting started.

console.log(JSON.stringify(project, null, 2))