💡 Tip: Press SPACE to mine faster!
Artificial Intelligence & Robotics Engineer
I'm a Software Engineer and a Master's student in Computer Science at Rutgers University (GPA 3.85/4.00). I design and ship full-stack systems end to end — APIs, pipelines, databases, and the deployment around them — with a specialty for putting AI, machine learning, and robotics to work inside real products.
I've built backends handling 150+ requests a minute at 98% accuracy, cut API response times by 40%, and engineered semantic search into a live government portal for the Government of India. This summer I wrote the autonomous navigation software for a 1,000-pound field robot at the U.S. Department of Agriculture — a full GPS-RTK + depth-vision pipeline on ROS2 that hit 90% autonomy across 10+ acres.
My stack runs from Python, FastAPI, and PostgreSQL to Docker, CI/CD, and ONNX — plus four peer-reviewed papers and a Rutgers-funded ML project along the way. I care about one thing above all: software that's not just clever, but reliable, scalable, and genuinely useful. Poke around the ⛏️ Projects and 💼 Experience tabs — there's a lot to dig up.
The languages, frameworks, and systems I build with — engineering first, AI-fluent throughout
Domain-specialized RAG system answering questions on SEC filings (10-K, 10-Q) and CFR Title 17 rules. Dual retrieval pipelines, query routing, evidence ranking, and hallucination guardrails.
Metadata-infused, attention-gated U-Net for lower-grade glioma segmentation. Fuses clinical metadata (age, histology, genomic clusters) with imaging, reaching 0.896 Dice / 0.814 IoU.
Rutgers-funded ML system predicting polymer energy states from 2,916 CIF structures. DFT-based features + stacked ensembles (RF, XGBoost, LightGBM, MLP) lifted performance from 73% to 80% MCC.
U-Net pipeline detecting drivable crop-row centerlines from Farm-ng robot RGB+Depth frames. Custom dataset, automated mask generation, and BCE/Dice training for real-time field navigation.
Executed black-box model inversion attacks on CNN models using autoencoders. Designed probability-rounding, noise injection, and Isolation Forest defenses.
Created behavior-based authentication systems checking keystrokes and login anomalies with Isolation Forest machine learning on 2500+ records.
Built big data preprocessors and interactive 3D dashboards displaying pre and post-pandemic employment projections from BLS statistics.
Analyzed deep learning and statistical forecasting models on multi-series datasets, hitting minimal MAPE (0.18) using N-BEATS network.
Created an end-to-end data pipeline: scraped jobs data, preprocessed outliers, and deployed models predicting salaries with 89.9% accuracy.
Designed a scalable product database and backend RESTful APIs with PostgreSQL. Integrated containerized testing and deployments on Docker.
My professional developer timeline
Certifications, metrics, and milestones
Specialization in networking infrastructure, routing switches, and network configuration safety pipelines.
Credly verifiedCompetencies in cloud computing architectures, load balancing, serverless systems, and cloud databases.
AWS CertifiedDatabase structure designs, optimized SQL queries, indexing metrics, and relational schemas configurations.
Verified CertificatePublished in CRC Press, Springer Nature, Springer India, and Elsevier on ML in education, textile AI, sentiment analysis, and time series forecasting.
View Publications ↗Peer-reviewed research published in Springer, Elsevier & CRC Press
A comprehensive overview of opportunities and challenges for applying machine learning across educational systems.
Explores how AI and machine learning are transforming modern textile manufacturing and production workflows.
A survey of NLP-driven sentiment analysis techniques and their application to forecasting stock market movements.
Benchmarks deep learning and statistical forecasting architectures across multiple time series datasets.
Let's build something together — reach out anytime!