Mohan Official

Data Scientist

SANVAN Software Limited

Sep 2024 - Present

Engineered automated data cleaning and preprocessing pipelines using Pandas & NumPy on datasets exceeding 80K+ records, reducing manual preprocessing time by ~40%. Conducted comprehensive EDA using Python (Matplotlib, Seaborn) and SQL to surface actionable trends, directly informing 3 product strategy decisions. Built and benchmarked ML classification models (Logistic Regression, Random Forest, XGBoost) achieving up to 91% AUC-ROC — improving prediction accuracy by 18% over baseline. Developed feature engineering techniques (encoding, scaling, dimensionality reduction) that improved model F1-score by 12% across multiple projects. Wrote optimised SQL queries against MySQL and MongoDB to extract, filter, and transform data from 5+ data sources for modelling pipelines. Delivered 4 interactive Power BI / Tableau dashboards tracking key business KPIs, enabling weekly data-driven decisions for stakeholders. Deployed and monitored ML models in test environments, cutting deployment turnaround time by 25% through automated evaluation scripts. Collaborated cross-functionally with data engineers and business analysts to align model outputs with business objectives, presenting findings to non-technical stakeholders. Documented data workflows, model architectures, and experiment results, creating a reusable knowledge base that onboarded 2 new team members.