Ayon Roy

Currently

Pursuing MS in Data Science

Columbia University, NY

AI Research Intern @ Qualcomm ( Jun 2026 - Sep 2026 )
🎓
Research Assistant & TA @ Columbia University ( Sep 2025 - Present )
🖥️
Executive Data Scientist @ NielsenIQ ( Aug 2021 - Aug 2025 )
🌍
Community Founder India's 1st Kaggle Days Meetup
📚
Book Reviewer 4 published technical books reviewed

I help machines
learn smarter

Hey there, I'm Ayon Roy, an AI Researcher and MS Data Science student at Columbia University. My work spans Edge AI Quantization (Qualcomm), Generative & Agentic AI Systems (Columbia), and Multi-Modal / Geospatial Machine Learning (NielsenIQ). I specialize in ONNX runtime optimization (AIMET-ONNX), LLM orchestration, Knowledge Graphs (AWS Neptune), and translating complex data into scalable intelligence.

I served as an Executive Data Scientist at NielsenIQ's Global Centre for Statistical Research and Data Science for 2 years. Prior to that, I served as a Data Scientist in NielsenIQ's India Data Science Business Leaders team working with the biggest FMCG players in the world.

I founded India's 1st & Largest Kaggle Days Meetup community, reviewed 4 published technical books, have mentored and judged 115+ hackathons, and delivered 100+ technical talks worldwide. Having helped 3,950+ people start their Machine Learning journey, I genuinely believe in the power of community learning. I remain dedicated to helping "artificial brains" reach their full potential through adaptive problem-solving and scalable, real-world data strategies.

How I Play with Data

My expertise spans the full data science pipeline — from problem framing to production deployment.

📈

Predictive Modelling

Regression (Linear, Logistic, Polynomial, Ridge, Lasso), ARIMA & LSTMs for time-series forecasting, ensemble methods.

👁️

Computer Vision

Object detection (YOLO), image segmentation, satellite imagery analysis using OpenCV, TensorFlow and PyTorch.

💬

NLP

Text preprocessing, feature extraction, similarity detection, sentiment analysis and transformer-based approaches.

🔧

Data Pipelines

End-to-end data gathering, cleaning, feature engineering and model deployment at scale for real customer impact.

📊

Analytics & Insights

EDA, business hypothesis testing, experimentation analysis and storytelling that drives decisions.

📄

Research

Reading, extracting and implementing insights from recent research papers. 4 technical books reviewed.

Looking for a Data Scientist?

If you have opportunities, crazy product ideas, or want to collaborate — reach out. I'd love to chat.