Engineering
Intelligence.
I construct the bridge between deep learning research and high-performance full-stack architectures. Extracting signals from noise, designing scalable systems.
System Profile
Academic & Background
I am Arnab Das, pursuing a B.S. in Computer Science and Data Analytics at IIT Patna (GPA 9.18).
From researching satellite image dehazing at IIT Ropar to building financial analytics platforms, my work oscillates between the precision of scientific computing and the creativity of software engineering. I engineer intelligent systems that solve complex problems with both mathematical rigor and intuitive user experience.
The Arsenal — Skills & Tools
Python
TensorFlow
React.js
Node.js
Java
MongoDBPythonTensorFlowReact.jsNode.jsJavaMongoDB
MySQL
GitHub
NumPy
VercelMySQLGitHubNumPyVercelProfessional Experience
Research Intern
Indian Institute of Technology Ropar
Conducting research on deep learning-based satellite image dehazing. Re-implementing MCAFNet on the RRSHID dataset. Resolving architectural ambiguities via cosine-diversity loss and Coordinate Attention-based fusion.
Data Analyst Intern
Bluestock Fintech
Built a full-stack financial analytics platform for top Nifty companies with an automated ETL pipeline computing key metrics. Developed interactive dashboards and real-time integrations deployed on serverless architecture.
AI Intern
Edunet Foundation (Microsoft Initiative)
Engineered "Smart Waste Segregator," an IoT-integrated AI system using a ResNet50 model in TensorFlow (92% accuracy) deployed on Azure ML. Built React.js dashboard for real-time monitoring.
Selected Works
View GitHubZ-PRIME ENGINE
Z-Prime Anomaly Engine
Designed a Deep Autoencoder architecture in TensorFlow/Keras to model latent representations of Standard Model background data. Achieved 50% Signal Efficiency with a background rejection factor of ~9.2.
Collision Hunt
Engineered a high-performance ML pipeline using Uproot and Awkward Array, bypassing legacy C++ dependencies. Deployed XGBoost classifiers for rare signal event discrimination.
LYNQIT
LynqIt Platform
Built a robust real-time messaging ecosystem using the MERN Stack and Socket.IO. Features include AI-powered auto-replies, JWT authentication, and latency optimization successfully handling 100+ concurrent users in a scalable architecture.