🧠 Architecting Intelligence, One Model at a Time.

Hi! I'm Debopam Ghosh, a passionate developer of Backend Systems and Machine Learning.

---

🛠️ The Arsenal

My tech stack for building robust APIs and predictive models:

🤖 Machine Learning & AI

⚙️ Backend & DevOps

---

💡 What I Bring to the Team

I don't just write code; In a hackathon I can manage:

• Architects APIs: Building fast, secure, and scalable endpoints using FastAPI/Flask.
• Trains Models: Turning raw data into actionable predictions using ML algorithms.
• Integrates AI: Connecting LLMs (GenAI) to real-world applications.

---

🏆 Key Projects

1. AI-Powered Stock Indicator

A predictive analysis tool for financial markets.

• Tech: Python, Machine Learning (LSTM/Regression), Flask Backend.
• Impact: Analyzes historical data to predict future stock trends with high accuracy.

2. Amazon ML Challenge 2025

A Multimodal Price Prediction Pipeline using Vision & NLP.

• Architecture: Built a hybrid model fusing ResNet50 (Image Embeddings) and TF-IDF (Text Features).
• Engineering: Implemented a memory-efficient Chunking Pipeline using Sparse Matrices to process 75k+ high-res images and text rows on limited RAM.
• Model: Optimized LightGBM Regressor with Hyperparameter Tuning (RandomizedSearchCV) to minimize RMSE.
• Tech: Python, PyTorch, LightGBM, Scikit-Learn, Pandas.

---

Let's build something intelligent! 🚀