Jul 2024 – Present · 1 yr 11 mos

• ML team lead for projects in TB and maternal health.
• Also involved in building internal cross-cutting software libraries, government and policy engagements, responsible AI, recruitment, capacity building and training.

May 2022 – Jul 2024 · 2 yrs 2 mos

• ML team lead for multiple projects in healthcare. Setting the technical direction of the team and actively engaging with diverse stakeholders to understand and implement requirements.
• Lead author on best paper at ML4Health for the TB adherence project: https://proceedings.mlr.press/v193/kulkarni22a/kulkarni22a.pdf
• Other activities: hiring, internal trainings, responsible AI.

Oct 2019 – Apr 2022 · 2 yrs 6 mos

• Working on applying AI and data science solutions to problems in the healthcare space at a Gates Foundation, USAID and Google.org-funded AI for Social Good non-profit.
• Main project: Models to predict proactively predict adverse outcomes tuberculosis, disease affecting 10 million individuals per year globally.
• Achieve 1.7x lift above best baselines, 2.8x above random on the Recall@K=20% metric.
• Designed and ran exhaustive evaluations on different cohorts and geographies
• Use gradient boosted decision trees and neural networks, with multiple encodings
• Actively contributed to aspects of calibration, interpretability, fairness and uncertainty
• Project successfully evaluated by India's Central TB Division, pilot being planned.
• Past: HIV adherence models, resource forecasting to assist the BMC in managing the COVID-19 pandemic.

Jun 2018 – Sep 2018 · 3 mos · Mountain View, California, United States

• Worked with the Juicer team within the Machine Intelligence organization to run experiments on terabytes of data, which involved modifying and improving a production machine learning and information extraction C++ pipeline on email data.

May 2015 – Jul 2015 · 2 mos · Bengaluru Area, India

• Developed a statistical model to predict customer response to email sequences, based on Rubin's causal model. Successfully filed a patent:
• <https://www.google.com/patents/US20170213236>

Apr 2015 – Mar 2016 · 11 mos · Mumbai Area, India

• Insight is IIT Bombay's official English student media body, with the largest audience of any, globally.
• Awarded the Institute Organizational Colour.

May 2014 – Jul 2014 · 2 mos · Madison, Wisconsin Area

• Interned in the Space Science and Engineering Center (SSEC). Wrote an application to convert NASA images to be suitable for use by Linux-based weather software.