Naveen Mathew Nathan
Data Scientist & ML Researcher
Hands-on leader grounded in foundational principles of AI and their practical implications in system and solution design
I'm a (data) scientist with 11 years of industry experience in machine learning, AI, and data science. I specialize in Agentic AI from first principles, combining deep theoretical understanding with practical system design and solution architecture. My work focuses on applying mathematics for innovative solution design, building scalable AI systems grounded in foundational AI principles while delivering tangible business value through cost-effective system design and strategic implementation.
Boston, MA
About Me
Background & Interests
In the past I worked with client delivery teams to create automated pipelines for supervised and unsupervised learning. My research interests span across:
Research Areas
- Unsupervised learning and crowd-sourcing for exoplanet candidate identification
- Reinforcement learning for traffic control (US patent approved)
- Information theory based redefinition of binary cross-entropy for GLM
- Applications of ML and deep learning in finance
Knowledge Sharing
I maintain two blogs to share knowledge:
- Statistical & ML Analysis Blog - Detailed exploration of mathematical and statistical fundamentals of machine learning and deep learning
- Industry Solutions Blog - High-level discussion of industry expertise and innovative solutions I've developed
I also maintain a machine learning book with my blog posts in a loosely ordered format.
Product & Solution Development
Check out my solutions blog for more articles. Short presentations on my R&D work in industry:
Automated Pipeline for Binary Classification
An end-to-end automated pipeline for binary classification with comprehensive model evaluation and ensemble techniques.
Automated Pipeline for Clustering & Basket Analysis
Automated clustering pipeline with basket analysis for unsupervised learning and segmentation tasks.
NLP & Predictive Modeling Pipeline
Pipeline for natural language understanding and predictive modeling with comprehensive text mining techniques.
Research & Publications
Check out my Stat/ML blog or my full ResearchGate profile for more. Featured research work:
Traffic Management for Unmanned Aircraft
Patent · Nov 2023US Patent (US11830371B2) co-invented at IBM Research with Mudhakar Srivatsa, Raghu Kiran Ganti, and Linsong Chu, covering systems for managing air traffic involving unmanned aircraft. This work is the foundation for my ongoing open-source project to automate and optimize air traffic control using reinforcement learning.
Automated Insulin Delivery Research
Medtronic · 2024–2026Real-world evidence studies on the MiniMed™ 780G automated insulin delivery system, analyzing glycemic outcomes across 40,000+ users — including nighttime burden reduction, sleep quality, and equity of outcomes across socioeconomic groups.
Deep Learning in Finance
Deep learning for predicting VWAP movements in limit order book data, plus multi-stage financial modeling in R presented at useR! 2020. Also includes work on scaling terabyte-scale financial data processing.
Exoplanet Candidate Identification
Unsupervised learning and crowd-sourcing approaches for identifying exoplanet candidates. Includes interactive Shiny application for candidate review.
Theoretical Work
Information theory based redefinition of binary cross-entropy with applications to generalized linear models.
Get in Touch
Feel free to reach out for collaborations or inquiries.