AJAY
Currently pursuing MS Business Analytics @ WPI
I work at the intersection of business and data, building solutions that turn raw information into clear decisions. With a background in business development and hands-on experience in analytics, I bring both context and execution to the table.
My work spans machine learning pipelines, recommendation systems, and marketing analytics using Python, SQL, and BI tools. I focus on making data not just accurate, but actionable for real-world decisions.
I am also the founder of DataStatz, a platform that simplifies analysis through automated insights, AutoML, and shareable dashboards.
Ajay Ramineni
Data & Business Analyst
0.0 GPA
Academic Excellence
0+
Leads Managed
0K+
Records Analyzed
0+
Projects Delivered
Selected Work
PROJECTS &
CASE STUDIES
Azure ML Income Prediction
End-to-end ML pipeline for income prediction using boosted decision trees on Azure ML Studio. Includes data preprocessing, feature engineering, hyperparameter tuning, and model deployment. Achieved strong AUC-ROC on the Adult UCI dataset.
// Key Highlights
- ▸Built full pipeline from raw data to deployed model on Azure ML Studio
- ▸Applied feature engineering, cross-validation & hyperparameter tuning
- ▸Achieved optimised AUC-ROC using ensemble boosted decision trees
Nike Business Intelligence Dashboard
Analyzed 3 interconnected datasets (99k+ records total) to surface regional sales KPIs, retailer performance, and product trends. Built interactive Power BI dashboards with drill-down capabilities. Proposed a unified data warehouse strategy for cross-team alignment.
// Key Highlights
- ▸Processed 99K+ records across 3 datasets to surface actionable KPIs
- ▸Built interactive Power BI dashboards with regional drill-down
- ▸Proposed unified data warehouse strategy for cross-team alignment
Anime Recommendation System
Content-based recommendation engine built on 12,000+ anime titles. Uses TF-IDF vectorization on genre, type, and studio metadata, with cosine similarity for personalized recommendations. Includes popularity weighting and cold-start handling.
// Key Highlights
- ▸Vectorised metadata across 12,000+ titles with TF-IDF
- ▸Personalises suggestions by genre, type, and rating signals
- ▸Includes popularity weighting for cold-start handling
Titanic Kaggle Competition
Kaggle survival prediction competition. Applied feature engineering (title extraction, family size, cabin deck), ensemble methods combining Random Forest and XGBoost, and k-fold cross-validation. Final submission achieved 0.78708 accuracy (Top 35%).
// Key Highlights
- ▸Feature engineering: title extraction, family size, cabin deck
- ▸Ensemble of Random Forest + XGBoost with k-fold cross-validation
- ▸Ranked Top 35% globally with 0.78708 test accuracy
Tech Stack
SKILLS & TOOLS
Frontend
Backend & APIs
Machine Learning
Data & Databases
Business Intelligence
Tools & Platforms
Writing
LATEST
ARTICLES
Get in Touch
OPEN TO DATA, ML & BI ROLES
Actively looking for opportunities in data analytics, business intelligence, and machine learning.