Puspendra Singh Email and Phone Number

1. Predicting diabetic Nephropathy: Developed a robust prediction model to identify the risk of diabeticnephropathy in advance, aiming to prevent kidney disease or failure. The model incorporated ECR and eGFR values and utilized two approaches: visit history and the RCS approach. Employed the XGBoost model for training and prediction and precision-recall curves for model performance.2. Atrial fibrillation disease detection: Contributed to the development of a primary care tool that effectivelyidentifies patients at risk of A-Fib. Improved the existing model by incorporating clinical input and ensuredflexibility for adaptation to various datasets.3. POC: Abstractive and extractive text summarization and pdf table extraction using the Python library.

1. Propensity of buying life/health insurance policy: Developed a classification-based machine learningmodel to predict customer likelihood of purchasing insurance. The model significantly improved customertargeting by identifying individuals with a high propensity to buy. Additionally, I implemented a rule-basedHI/LI monthly trigger to engage customers who had previously purchased HI/LI in the same period.2. Built a rule-based income estimation model based on historical data patterns, effectively establishing acorrelation between features and salary.3. Implemented a customer segmentation model based on demographics and transactional behaviours,utilizing unsupervised clustering algorithms.

1. Project: Built robust Univariate/Multivariate anomaly and outlier detection model using statistical, Machine learning, and deep learning algorithms like Holt’s Winter, Luminal, ADTK, Twitter, Prophet, LSTM, etc. at device level/Counter group/node/KPI level category on raw/hourly/daily data. 2. Developed a baseline deviation threshold using statistical algorithms to set alarms for abnormal data.Conducted exploratory data analysis to identify trends, seasonality, and other insights on raw, hourly ordaily data. Built an ML-based time series forecasting model and trend analysis to enhance telecom networkperformance.3. Project: Developed an ML-based classification test suite optimization model to minimize redundant testing and automate test suite selection. Leveraged NLP (Regression expression) to extract data from PDF release notes.

1. Project: Conducted credit card default analysis and prediction of delinquent customers. Performedexploratory data analysis, data cleaning, missing value treatment, and outlier treatment2. Utilized Chi-square, ANOVA tests, hypothesis testing, and graphical analysis to determine the significance of features with respect to the target variable. Applied feature engineering techniques such as binning, feature creation, and conversion of variables into Dummy variables.3. Developed a response model at the customer level using logistic regression, boosting, and baggingtechniques to enhance algorithm accuracy.

1. Utilized regression models to predict sales for retail customers based on various features. Additionally, also utilized forecasting algorithms to project sales using time-dependent data.2. POC: Build an ML model using Association rule and collaborative filtering to discover which groups of products tend to be purchased together. This analysis will help us understand customer purchasing behaviour and enhance cross-selling opportunities.3. Prepared revenue sales projections, MIS reporting, ad-hoc reporting, productivity analysis, EBITDA andrevenue analysis, outstanding reports, and other sales reports.

1. Created MLS, KIT sheets, method models, tool orders, VWI, and SAP cashier for each part. Modified partmodels using Catia-v5 and 3D-via software. Contributed to the delivery of major parts including ForwardFuselage, Stabilizer, Nose Skin, Wings, and Tail Cone.2. Coordinated on-site and offshore teams of 15 individuals, standardizing shop floor processes. Significantly improved shop floor on-time delivery and offshore project quality.