Aarin Singh

Created a framework to streamline the projects for Azure ML studio. Extracted data for ad-hoc data requirements and built dashboards on Metabase.

Constructed user’s feed recommendation classifier powered by Python, AWS Sagemaker, and XGBoost with 70% accuracy; resulting in a 300% boost in the engagement matrix. Forecasted non-app launch users by utilizing Python, XGBoost with 75% recall rate, leading to a substantial 25% upswing in app launch and 38% in quality sessions. Implemented a personalized notification system using Python, Scikit-learn, and Kmean algorithm, yielded a remarkable 30% increasing in CTR , and the consequential impact was a staggering 280% surge in the click-to-engagement ratio. Developed Multi‑Class Topic Tagging Model using Fast.ai and Python, achieving 72% accuracy, ultimately saving over 8‑10 hours of manual work daily. Conducted performance analysis by utilizing python libraries such as pandas, matplotlib, and seaborn to replace underperforming features, resulting in 10% less cost on production and an increment of 30% in user session time. Successfully onboarded leading brands like Sonalika, Escorts, Stihl, MRF Tyres, Corteva, and Bayers on Krishify by executing qualitative analysis on e-commerce keywords. Created dashboards as per stakeholders’ requirements from databases like BigQuery, RDS, and Snowflake; published it on Metabase.

Analyzed large amounts of information to discover trends and patterns. Identified KPIs and maintained KPIs reports to keep track of key matrices. Identified cohorts to classify the types of users to introduce subscriptions resulting in 35% more subscribers. Performed diagnostic analysis of features to identify hooks of Corploud which resulted in an increment of users session numbers in a week by 15%.