Kashish Thakur

I’m a data analytics and machine learning professional with an academic foundation in Data Analytics (San Jose State University) and Computer Science (Lovely Professional University). I bring a strong emphasis on end-to-end delivery—from data preparation and modeling to clear outputs that stakeholders can use.

In my projects, I focus on practical, measurable impact. I built an LLM-Driven Automated Design-to-Code System using NLP, RAG, and Generative AI (GPT-4o and other models), reducing design-developer handoff efforts while improving design accuracy (F-Score 0.53, BERTScore-F1 0.59).

I also develop models for real business problems and validate them with strong performance. For Subscription Churn Rate Prediction, I used ML algorithms (Random Forest, SVM) and achieved 94% prediction accuracy, supported by extensive EDA and 15 insightful visualizations to enable proactive retention strategies.

Before concentrating full-time on advanced analytics work, I served as an Assistant Professor, designing and delivering labs and lectures that improved outcomes—boosting student engagement and retention by 40% and raising course completion rates by 10%. I enjoy building with structured tooling (Jira, Trello) and translating complex technical work into results people can act on.