Junior Research Fellow at NIT Raipur, working on cybersecurity projects with a focus on Graph Neural Networks and Deep learning. Skilled in Python and PyTorch, and passionate about exploring AI to solve real-world problems
-
🔭 I'm currently working on Utilizing variants of Graph neural network for intrusion detection models for advanced persistent threats.
-
🌱 I'm currently learning Graph neural networks, RAG based LLM's, Agentic AI systems
-
📫 How to reach me [email protected]
-
📄 Resume Drive Link
- Tech Stack: Astronomer Airflow, Redis, PostgreSQL, Prometheus, Grafana, Docker, GCP.
- Built a comprehensive MLOps pipeline for Titanic survival prediction with end-to-end automation.
- Implemented data drift detection using Alibi Detect for model monitoring.
- Orchestrated ETL pipelines using Astronomer Airflow with GCP integration.
- Created Redis-based feature store for real-time feature caching.
- Set up comprehensive monitoring with Prometheus and Grafana dashboards.
- Containerized deployment with Docker and automated CI/CD workflows.
- Built complete MLOps pipeline with Jenkins CI/CD and GCP CLoud Run deployment.
- Implemented automated model training, validation, and deployment workflows.
- Containerized application for scalable production deployment.
- Developed personalized recommendation engine using collaborative filtering and content-based methods.
- Implemented matrix factorization techniques and similarity metrics.
- Created interactive interface for anime discovery based on viewing preferences.
- Built complete MLOps pipeline with Jenkins CI/CD and GCP GKE deployment.
- Tech Stack: FastAPI, Python, Gemma AI, JavaScript, Docker, Railway
- Developed intelligent research assistant for searching and chatting about arXiv papers
- Integrated FREE Gemma-3n-e4b-it model for unlimited AI conversations
- Built real-time paper search with advanced filtering and metadata display
- Implemented context-aware explanations combining paper content with AI knowledge
- Features include bookmarking, search history, conversation export, and mobile responsiveness
- Deployed on Railway with automated GitHub integration
- Tech Stack: HTML, JavaScript, CSS
- TextCraftAI is a modern text processing web application that leverages transformer-based AI models (PEGASUS and T5) to provide powerful text summarization and paraphrasing capabilities.
- The application offers a clean web interface and robust API endpoints to process plain text and various document formats.
- Tech Stack: Python, Flask, Docker, Kubernetes, HuggingFace
- Built containerized sentiment analysis application
- Implemented deployment pipeline with Docker and Kubernetes orchestration
- Tech Stack: PyTorch, CNN, Weights & Biases
- Built CNN-based classifier for food vs. non-food detection using custom curated ImageNet-1k dataset
- Achieved 93.67% training accuracy and 87.28% validation accuracy
- Enhanced model stability through Batch Normalization and Dropout techniques
- Used Weights & Biases for real-time training monitoring and optimization
- Tech Stack: TensorFlow, Keras, CNN
- Developed high-accuracy binary classification model for cat and dog image identification
- Enhanced performance through data augmentation techniques (rotation, flipping)
- Optimized hyperparameters to improve training efficiency
- Implemented Graph Convolutional Network for terrorist attack pattern classification
- Applied graph-based learning to cybersecurity and threat detection scenarios
- Tech Stack: Python, Pandas, Scikit-learn
- Built predictive model using advanced feature engineering (frequency encoding, missing value handling)
- Explored multiple algorithms: Linear Regression, Random Forest, Gradient Boosting
- Achieved significant improvement in price prediction accuracy through hyperparameter tuning
- Tech Stack: Python, Pandas
- Comprehensive analysis of US bike share data using statistical methods
- Data exploration and visualization for transportation pattern insights
- Advanced GNN Architectures: Exploring novel graph neural network variants for cybersecurity applications, specifically for intrusion detection in advanced persistent threats
- RAG Systems: Building sophisticated retrieval-augmented generation systems for domain-specific applications with arXiv integration
- MLOps Pipelines: Developing robust machine learning operations workflows with comprehensive monitoring and drift detection
- Agentic AI Research: Investigating autonomous AI agents for complex reasoning and decision-making tasks in cybersecurity contexts
- Graph Neural Networks for Cybersecurity
- Advanced Persistent Threat Detection
- MLOps and Production ML Systems
- Retrieval-Augmented Generation (RAG)
- Autonomous AI Agents
- Deep Learning for Computer Vision