At the Dept. of Mechanical Engineering, MNNIT Allahabad, I developed a machine learning pipeline to predict eight bearing performance parameters from 152 tribology records. The FFNN model achieved an average test R of 0.991 and MAPE of 2.6%.
I also deployed a Streamlit app with Plotly visualizations for model comparison, parametric sweeps, and validation against published data.
On DeliverySense, I designed a relational schema and Python ETL pipeline to load 100K+ orders into PostgreSQL. I used SQL root-cause queries to examine delivery delays and their impact on review scores, then launched a Streamlit dashboard.
For my AI Tool Adoption Analytics Dashboard, I prepared 50K+ records, created Power BI measures and interactive visuals, and trained a baseline linear regression model to test adoption-rate predictors.
