NIKHIL SACHDEVA

Replicated statistical analysis for LVEF FDA Submission including model subgroup analysis and sample size for clinical validation; identified critical gaps and facilitated ownership of further FDA approval studies. Conducted experiments establishing link between ECG model predictions for multiple cardiovascular diseases and long-Covid, laying foundation for predictability of long-Covid in patients. Worked on the ASCVD disease diagnosis model including creation of stenosis, ischemia and Akinesis cohorts, model training and performance analysis for FDA submission and utility analysis for Novartis, dealing with the challenges of combining multiple disease models under a single umbrella. Provided statistical expertise to various teams for their projects, including ECG Clustering Analysis, Lead Reversal Experiments and Inferential Statistics Classes. Received Bravo Award twice for exceptional performance. Performed 2020 VBM Testing to substantiate the performance of VBM Model including statistical testing related to frequency of VaR exceptions and distribution tests to check conservativeness of VaR. Worked on VaR reviews including RMBS(Residential Mortgage Backed Securities), Repo rates and FX Basis. Prepared analytical background for automation of Peer Bank comparison of Assets and Liabilities using OCR functionalities. Implemented ULMFiT (Universal Language Model fine tuning) to identify SCRA applicability from call transcripts due to small training data size and its generalizability to other disclosures.

Courses: Measure Theory, Statistical Inference, Time Series Analysis

Courses: Regression Techniques, Probability Theory, Optimization