Jayant Apte, Phd Email & Phone Number

Charlotte, North Carolina, United States

• Contributed to various types of projects including client facing analytics consulting, ML based enhancements to existing algorithmic underwriting product, and applied research in the areas of survival ranking, knowledge distillation, and group fairness • Part of the team that modernized existing algorithmic underwriting product by integrating several survival ranking and binary classification models, which led to the client entrusting the employer with additional reinsurance share (equivalent to several million dollars in additional annual premium)• Managed 3 interns over 3 summers, provided hands-on coaching on topics such as python, machine learning, insurance and risk data science - 1 intern was hired to be full-time employee with the team, 1 intern was hired internally by another team, 1 intern is on their way to publish a manuscript based on the research they conducted with me during the summer• Lead projects that addressed business pain points by using deep analytics dashboards that delivered decision support metrics e.g. anomalous client portfolio detection in property insurance to speed up assessments of client portfolios during renewal season and, quantification of dollar value of vendor data consumed by a life insurance underwriting engine to support underwriting strategy changes• Created and conducted several team-wide and organization-wide trainings on topics such as generative AI, python (basic and SOLID design patterns), storytelling with data (based on the eponymous book), and knowledge sharing regarding ML driven product enhancements

Charlotte, North Carolina Area

• Trained and validated xgboost cox proportional hazards models with to predict relative risk score for life insurance underwriting application and cohort level Breslow estimators for predicting expected number of events for pricing applications, assisted machine learning engineers in productionizing the models (wrapper code, model versioning etc)• Collaborated with medical underwriters to understand model decisions using SHAP local explana-tions and dependence plots, experimented with free parameters of model training process to smooth out any anamolies• Trained xgboost binary classification models for evidence waiver, conducted detailed reproducible experiments to support data driven choices and understanding of bottom-line financial impact of incorporating a model in the underwriting workflow• Developed robust data science practices around experiment tracking, data and model artifact version control on top of gitflow that were adopted within the agile framework as definition of done for the team

Greater Philadelphia Area

• Collaborated with a team of physicians, engineers, and data scientists in a specialty consulting startup environment, eventually leading multiple client-facing projects simultaneously• Worked with clients and datasets from diverse industries/sectors viz pharmaceutical, health insurance, auto parts, non-profit for veterans, media/publishing• Co-developed an auto-ml pipeline to diagnose undiagnosed rare disease patients from US population scale health-insurance claims databases (ICD/CPT/NDC coded temporal data)• Worked with a manufacturer of biologic drugs to compute metrics such as market size, incidence/prevalence, healthcare provider lists to target with educational material from EHR and health insurance claims data, all of which informed their go-to-market strategy for a drug• Worked with a global manufacturer of vaccines to create a database containing vaccine regulatory landscape information in ~130 countries via web scraping, which was used for identifying key opinion leaders in foreign markets using analysis of network of medical publication authors• Worked with a strategic communications client to generate insights around product trends from social media firehose data using NLP

Philadelphia

Implementing low-complexity algorithms that perform class membership testing, extension and visualization of matroids. Using matroid extension to determine fundamental limits on transfer and storage of information in networks.

Drexel University

Teaching assistant for the course ECE-C413 Parallel Computer Architecture

Edmund D. Bossone Research Center

I worked as a Teaching Assistant for the course ECES-436 Multidisciplinary DSP. My responsibilities included helping students through the lab sessions and grading them for the same.

Doctor Of Philosophy (Ph.D.), Electrical Engineering

Activities and Societies: IEEE

Master Of Science - Ms, Electrical Engineering

Activities and Societies: IEEE

Bachelor Of Engineering (B.E.), Electronics And Telecommunication Engineering