Diptarka Saha

** Was responsible for conducting discussion sessions, grading, and managing logistics for several in-person and online graduate-level courses with 300+ students. ** Instructed an undergraduate-level course with 50+ students. ** Ranked Excellent by their Students

** Subgroup Identification for Heterogenous Treatment Effect of Malaria Drug ** We devised a Meta-Learning based approach to quantify and infer subgroup effects on treatment. We identified interesting subgroups which respond heterogeneously to treatment. Methods/Skills used: Meta Learning, Xgboost, Neural Network, Rshiny, LASSO

** Forecasting Harvest Readiness using High-Resolution Satellite Data ** We used satellite Images & ML to develop an automated algorithm to forecast harvest readiness by capturing the naturally occurring temporal curves at a sub-field level ML Methods used: Xgboost, Generalised Additive Mixed Modeling

******** Work Experience ******** 1. Developed, published, and patented a novel clustering and anomaly detection algorithm – REDCLAN: RElative Density-based CLustering and ANomaly Detection. 2. Developed, published, scaled, patented, and helped in productionizing novel algorithms for determining Promotion Adjusted Loyalty of products from the customer point of sale data. 3. Developed, published, and patented automated stepwise approach to estimate optimal parameters in SARIMA models. 4. Developed a novel statistically efficient technique to estimate lost sales from observed sales and inventory data. 5. Developed an automated forecast engine that provides true demand for the user-defined price, promotion, and other input variables in the domain of apparel. ***** Awards and Recognition ******* • Was awarded Best Paper at National Conference on Machine Learning and Artificial Intelligence (NCMLAI), 2018 • Was awarded Associate of the Month (July 2018) among 2000+ associates of WalmartLabs (International)

Was part of the team that organized the placement process overseeing full-time and internship placements for several M.Stat, MSQE, and M.Tech candidates of ISI, Kolkata through the years 2016 and 2017

Developed algorithms to find new and hidden patterns and detected anomalies based on those patterns in financial data using several automated unsupervised learning processes

A classification problem: an approach to scorecard modeling for customer attrition data under the assumption of heterogeneous population using logistic regression, k nearest neighbour algorithm and combination of both

Thesis : 'DEEP MODELS, LIGHTER FOOTPRINT COMPRESSING, EXPLAINING, AND TRANSFERRING BAYESIAN NEURAL NETWORKS' Passed Courses: - Advanced Regression Analysis: Grade A+ - Probability and Measure: Grade A - Mathematical Statistics: Grade A - Mathematical Statistics II: Grade A+ - Probability and Measure II: Grade A - Mathematical Statistics: Grade A - Analysis of Variance: Grade A+ - Categorical Data Analysis: Grade A+ Recognitions: - Qualifying Exam Award - Elected : Member of Teaching Assistant Welfare Committee

- Passed the prestigious M.Stat course with first-class, with specialization in quantitative finance. - Final Thesis: Developing benchmarks and indicators for the creation of disaster risk (DRisI) index for districts of states and union territories of India

- Passed the prestigious B.Stat course with first class - Final year project: A classification problem: a twofold logistic regression-based approach to credit scoring method using quarterly binning of data