Deepesh Singh deepesh1singh

About Me

Indian Institute of Technology Kharagpur

Premier Technical Institute | Est. 1951 | Top 5 IIT

Pursuing an integrated B.Tech and M.Tech in Computer Science and Engineering at IIT Kharagpur. Solid expertise in data structures and algorithms, web development, and applied machine learning, with a passion for tackling challenging problems.

Featured Projects Showcase

Machine Learning-Based Differential Cryptanalysis of ASCON Permutation

Advanced machine learning approach to cryptanalysis of the ASCON permutation

Technical Details & Performance

Research Focus:
  ├── Target: ASCON permutation (lightweight cryptography)
  ├── Approach: ML-based differential distinguishers
  ├── Models: CNN, LSTM, LightGBM
  └── Dataset: 100,000 samples (50% random, 50% real)

Data Preprocessing:
  ├── CNN/LSTM: Decimal to 3D binary vectors
  └── LightGBM: Decimal to 2D binary matrices

Performance Metrics:
  ├── Round 1: 100% accuracy (CNN, LSTM, LGBM)
  ├── Round 2: 100% accuracy (CNN, LSTM, LGBM)
  ├── Round 3: 94.52% accuracy (LGBM)
  └── Rounds 4-5: ~50% accuracy (approaching random)

Key Achievements:
  ├── Successful detection of non-random patterns
  ├── Multi-model comparative analysis
  └── Identification of distinguisher limits

Tech Stack:

BankBrain – Cloud-Ready Microservices Banking Backend

Cloud-native banking backend with MCP-driven multi-agent architecture

Technical Architecture & Features

Architecture:
  ├── Pattern: Microservices
  ├── Framework: FastAPI (Python)
  ├── Containerization: Docker
  └── Orchestration: Kubernetes-ready

Multi-Agent System (MCP):
  ├── Support Agent: Customer assistance
  ├── A2A Gateway: Account-to-account transfers
  └── Risk Agents: Transaction monitoring

Testing Framework:
  ├── Type: In-memory end-to-end smoke tests
  ├── Tool: FastAPI TestClient
  └── Coverage: Complete workflow validation

Integration Features:
  ├── Mock Banking APIs
  ├── Tool-calling workflows
  └── Secure transaction retrieval

Key Features:
  ├── Modular design for scalability
  ├── Cloud-ready deployment
  ├── Comprehensive testing suite
  └── Secure transaction processing

Tech Stack:

Multimedia Database System

Full-stack multimedia content management system with real-time communication

Technical Architecture & Features

Supported Media Types:
  ├── Books
  ├── Videos
  ├── Music
  ├── Images
  ├── Articles
  ├── Newspapers
  └── Games

Security Features:
  ├── bcrypt: Password hashing
  ├── JWT: Token-based authentication
  ├── Helmet: Security headers
  ├── CORS: Cross-origin resource sharing
  ├── Rate Limiting: Request throttling
  └── Input Validation: Data sanitization

Database Architecture:
  ├── ORM: Sequelize
  ├── Database: SQLite
  └── Features: Efficient queries & indexing

Real-time Features:
  ├── Socket.IO: Live communication
  ├── Automated metadata extraction
  └── File optimization for efficient storage

Frontend:
  └── Bootstrap: Responsive design

Tech Stack:

Event Management System

Enterprise-grade event management platform with role-based access control

Technical Architecture & Features

Architecture:
  ├── Pattern: MVC (Model-View-Controller)
  ├── API Design: RESTful endpoints
  └── Database: MongoDB (NoSQL)

User Roles:
  ├── Manager: Full system access
  ├── Vendor: Service management
  ├── Accountant: Financial oversight
  └── Customer: Event participation

Core Features:
  ├── Event creation & management
  ├── Ticket booking system
  ├── Payment processing
  ├── PDF ticket generation
  └── Automated email notifications

Security Implementation:
  ├── JWT: Token-based authorization
  ├── bcrypt: Secure password hashing
  ├── CORS: Cross-origin protection
  ├── dotenv: Environment configuration
  └── Structured access control

Additional Features:
  ├── Reporting & analytics
  ├── User management workflows
  └── Modular architecture for scalability