Profile banner
Ayush Lodha

Ayush Lodha

Computational data scientist focused on climate and statistical modeling

Indianapolis, Indiana
iu
Joined February 2026
Favicon imagegithub.com
Favicon imagemedium.com
Favicon imagewayup.com

Network

7.1K connections
🇮🇳
IIT Gandhinagar Network
🌱
Climate Sustainability Leaders
🎓
University Researchers
🤖
AI/ML Data Scientists
📊
Product Operations Leaders
🌍
Climatebase Fellows
☁️
...

Summary

Computational data science practitioner with formal graduate training: Ayush completed a Master's in Computational Data Science at Purdue School of Science (IUPUI) and served in multiple research and teaching assistant roles at the School of Science, developing production-capable tooling (for example, a FastAPI-based data exchange server) to integrate computational models across languages and HPC environments. iu+2
Applied statistics and Bayesian modeling focus: Ayush authors technical write-ups on Bayesian approaches to classification and documents MCMC model-building workflows and diagnostics, demonstrating an emphasis on principled statistical modeling. medium
Climate-focused data work and professional development: Ayush participated in the Climatebase Fellowship (Cohort 6) and holds data science roles tied to climate-focused organizations, indicating engagement with climate literacy and climate-oriented data applications. credsverse+1
Active open-source contributor and educator: Maintains multiple public GitHub repositories (project code, learning materials) and has experience as a teaching assistant, signaling both a commitment to open-source practice and to sharing knowledge with others. github+1

Work

Education

Projects

Writing

I have been following out various articles for making RAG better, advanced and improving it's performance. I like the flow of the article and all the ideas at one place collected for ease of

July 22, 2024

Medium blog post commenting on Retrieval-Augmented Generation (RAG) practices and resources, reflecting interest in improving RAG performance and summarizing relevant ideas for practitioners.

Favicon imagemedium.com

Bayesian Way of Building Classification Models.

December 31, 2022

Medium article describing a Bayesian approach to binary classification problems (case study: car insurance claim acceptance), covering data preprocessing, choice of priors (normal and student_t), MCMC settings with rstanarm, model diagnostics, and comparative discussion with logistic regression baselines.

Favicon imagemedium.com

Hobbies

Technical blogging and writing about machine learning, Bayesian methods, and retrieval-augmented generation. medium