Deep learning, agentic AI, and cloud-native data pipelines — from research prototype to production deployment.
📍 New York, US · rnatarajanma@binghamton.edu · +1 607 370 8105
Full-stack ML expertise — from model research to production infrastructure.
End-to-end model development, computer vision, NLP, and hybrid quantum-classical models for research and production use cases.
Classical ML pipelines, feature engineering, ensemble methods, and unsupervised learning at scale.
Real-time object detection, image classification, and video processing for manufacturing and medical applications.
Production deployments on AWS and Azure with container orchestration, CI/CD pipelines, and experiment tracking.
ETL pipelines, SQL data modeling, and streaming infrastructure designed for reliability and scale.
Interactive dashboards and data storytelling that put analytics in decision-makers' hands.
Real systems I designed, built, and deployed — not tutorial clones.
End-to-end Retrieval-Augmented Generation system that answers questions over private PDFs using agentic routing between local knowledge, web search, and hybrid retrieval. Features grounded responses with citations, semantic search, and modular LLM orchestration.
Scalable NLP data pipeline for preparing large-scale text corpora for LLM training. Includes ingestion, cleaning, deduplication, quality filtering, tokenization, and reproducible preprocessing workflows for high-quality datasets.
AI-powered cybersecurity dashboard that combines Snort IDS logs with six unsupervised machine learning algorithms to detect anomalies and suspicious network behavior in real time through interactive visualizations.
AI-powered fitness companion that delivers personalized workout tracking, injury-risk prediction, recovery scoring, and nutrition recommendations using predictive analytics and intelligent health monitoring.
Research-oriented hybrid quantum-classical deep learning system that predicts blood groups from fingerprint images using CNNs integrated with PennyLane quantum circuits for enhanced feature learning.
Educational cybersecurity platform demonstrating Differential Power Analysis (DPA), Simple Power Analysis (SPA), kleptographic RSA attacks, and Shor's algorithm through interactive Streamlit visualizations.
Interactive Tableau dashboards analyzing airline passenger satisfaction across demographics, travel classes, delays, and service quality to uncover business insights through visual analytics.
Machine learning web application built with Flask that predicts loan approval outcomes based on applicant information using a trained classification model and an interactive prediction interface.
Internships building real ML systems across healthcare, telecom, and growth analytics.
Open to ML engineering, data science, and data engineering roles — open anywhere in the US.