Data Scientist and AI/ML Engineer specializing in production-grade machine learning systems, anomaly detection, and applied AI research. MCA graduate from VIT Bhopal (2024) with hands-on experience building end-to-end pipelines across cybersecurity, fraud detection, and customer intelligence domains.

Research focus: hybrid quantum-classical machine learning frameworks for real-time intrusion detection on NISQ hardware — with active work at the intersection of quantum computing and applied cybersecurity.

• International Best Researcher Award — Scopus Index Conclave 2025, recognized for applied AI research in cybersecurity and anomaly detection
• Designed and benchmarked a hybrid Quantum-Classical Intrusion Detection System (Q-IDS) using Variational Quantum Circuits on the UNSW-NB15 dataset
• Built a production-ready React + FastAPI demo for a live quantum-classical IDS, integrating QSVM, SVM, and Random Forest in a weighted fusion model
• Delivered fraud pattern identification on large-scale federal spending data (USAspending.gov), processing multi-GB datasets with distributed computing pipelines
• Consistent focus on translating research prototypes into deployable, measurable systems

Domain expertise: Anomaly Detection, Classification, Time Series Forecasting, Customer Segmentation, NLP, Quantum Machine Learning (VQC, QSVM, quantum kernels)

AWS services in active use: SageMaker, EC2, S3

Hybrid ML framework combining QSVM, classical SVM, and Random Forest for zero-day network intrusion detection

• Implemented Variational Quantum Circuits (VQC) on NISQ-compatible hardware using PennyLane, benchmarked against classical baselines on the UNSW-NB15 dataset
• Designed a weighted fusion model combining quantum and classical model outputs, resolving decision boundary bias and label-inversion issues in probabilistic inference
• Deployed as a React + FastAPI application with real-time prediction, mode-aware demo routing, and MD5-based feature signature caching for inference reliability
• Core contribution to MCA thesis: "Quantum Circuit-Optimised Framework for Zero-Day Intrusion and Anomaly Detection on NISQ Hardware"

Multi-layer identity protection platform combining behavioral biometrics and AI-based anomaly detection

• Designed an authentication pipeline using AI-based anomaly scoring to flag credential misuse and account takeover attempts in real time
• Integrated quantum-resistant cryptographic concepts as a forward-looking security layer
• Focused on reducing false-positive authentication blocks while maintaining high sensitivity to anomalous login patterns

End-to-end ML pipeline for CLV estimation in a retail/e-commerce context

• Engineered domain-relevant features from transactional data; trained and compared LightGBM and XGBoost regression models with cross-validated hyperparameter tuning
• Delivered a reproducible pipeline from raw data ingestion through model evaluation and prediction export, built for straightforward production integration

Anomaly detection on large-scale U.S. government spending data

• Processed multi-GB public datasets from USAspending.gov using distributed Hadoop pipelines, enabling analysis at a scale that ruled out single-machine approaches
• Applied unsupervised anomaly detection methods to identify statistical outliers indicative of procurement irregularities
• Produced interpretable findings suitable for audit review, not just model outputs

• Quantum ML research — advancing the Q-IDS framework toward publication-ready benchmarks on real NISQ hardware
• Production ML engineering — tightening the gap between research prototypes and deployment-ready, observable systems
• Algorithms and system design — reinforcing DSA foundations to support both competitive engineering roles and research implementation work

This profile reflects active work. Projects and research outputs are updated as they reach shareable milestones.