About
I'm Kharanshu Naghera, a machine learning engineer in Toronto. At RBC I build and ship models for credit scoring and fraud detection, plus the pipelines and serving code around them. That work runs on Azure (a risk co-pilot, drift monitoring) and GCP (a compliance platform built on LangGraph with ADK agents and MCP tools). Before that I spent two years at Apex Infolab in Bengaluru, where my recommender, NLP pipelines and AWS data platform served two million monthly users.
Each project on this site reports a held-out number against a baseline and the leakage check behind it. The eBay model ships an 80% band with every prediction, and its serving code exits when outputs drift more than 0.5% from a recorded reference.
Education and certifications
2023 - 2025
Trent University, Peterborough, ON
Masters in big data analytics · GPA 3.7 / 4.0
2018 - 2022
Gujarat Technological University, India
Bachelors in computer engineering
certificates
Google Data Analytics · AWS Cloud Management Console
More work
Global power plant analysis
SQL and Power BI over 34,936 plants in 167 countries: capacity by fuel and by country, and where the reporting gaps are.
Domestic violence EDA
Alluvial, dot-plot, and heatmap views of a domestic violence dataset; education, employment, and income against reported cases.
US baby names EDA
144 years of Social Security name records in MongoDB; trends, one-hit wonders, and the years the ratio flipped.
COVID-19 social distance analyzer
Computer vision on street footage, measuring distances between people; runs as a Colab notebook.
CouchDB vs MongoDB
Six experiments on FAO crop data, re-measured in 2026 and played back at the speed they happened; also on the home page.