About
Data Scientist and Analyst with a PG Diploma in Data Science & AI and 6+ years of experience in operations, including 3+ years as a team leader analysing CRM data and tracking KPIs. Built and deployed two live machine learning applications (FastAPI on Render, Streamlit Cloud) and currently work as a Data Scientist Apprentice.
Experience3
Data Scientist Apprentice
Jun 2026 — PresentSpinnaker Analytics · Internship · Remote
Work on the full workflow of ML projects including data cleaning, EDA, missing-value imputation, model training and evaluation. Build and compare classification, tree-based ensemble and time-series forecasting models in Python. Create Power BI dashboards that turn model outputs into decisions for business teams.
Team Leader – Operations
Jun 2022 — Nov 2025Hansa Direct, Quess Corp, Digitide Solutions
Analysed large CRM datasets to find productivity gaps and prepared MIS reports for senior management. Set and tracked SLAs, conversion and cost-per-lead targets for a team of up to 15 members. Used performance data to plan coaching and training, which improved team conversion metrics.
Operations Executive
Dec 2018 — May 2022Hansa Direct
Handled the Twitter and email support channels for the Mahindra Automotives account, resolving customer queries and complaints.
Projects3
Employee Attrition Prediction API
Built a job-change risk classifier on 19,158 employee records and deployed it as a live REST API with Swagger docs (POST /predict). Applied SMOTE and scaling on the training split only to avoid leakage. Compared four models; LightGBM was best with 79.4% accuracy, 0.812 ROC-AUC and 69% recall on the job-change class.
Open on github.com- FastAPI
- LightGBM
- SMOTE
- Render
Solar PV Annual Energy Estimation
Cleaned 218,115 records to 133,535 rows (11 features) and trained regression models on a log-transformed target. Random Forest reached R-squared 0.9998 and MAPE 0.08% on a held-out test set (5-fold CV R-squared 0.9996). Deployed on Streamlit Cloud and exported results to a Power BI dashboard.
Open on github.com- Scikit-Learn
- Streamlit
- Power BI
Hotel Booking Cancellation Prediction
Analysed 36,275 bookings (32.8% cancelled); cancelled bookings had a much longer lead time (139 vs 59 days). Tuned Random Forest over 108 configurations with 5-fold CV: 90.5% accuracy, 0.957 ROC-AUC. Built a Power BI dashboard on cancellation patterns.
Open on github.com- Random Forest
- GridSearchCV
- Power BI
Education2
PG Diploma in Data Science & Artificial Intelligence
2025 — 2026Boston Institute of Analytics
Bachelor of Computer Applications (BCA)
2021 — 2025Purvanchal University, Jaunpur
Certifications3
Machine Learning Algorithms
LinkedIn Learning
Data Science Foundations
LinkedIn Learning
Workshop: AI Dashboards using Power BI (2 days, live)
Skill Nation