Responsible for designing, building and maintaining backend services (Go / gRPC / Protobuf) and AI / LLM infrastructure and pipelines (Iceberg, BigQuery, Airflow, Docker, Label Studio) for medical code (CPT / ICD) generation and evals to enhance Suki’s clinical intelligence capability

As a Staff Engineer in the Data Platform Engineering team, I redesigned the data architecture to improve data governance, security and adoption and led Operational Excellence effort focused on data platform / pipelines observability improving data freshness, compute cost, data quality metrics and incident resolution SLAs.

An an EM, led a team of 5+ engineers and drove technical rigor through design reviews, observability dashboards, and incident postmortems. Guided the team towards reducing compute cost and evolving core data models while retaining and growing a group of talented engineers.

As an IC, worked on data platform and pipelines, core backend services focused on property data ingestion + risk scoring, and production grade ML pipelines to build the data foundation necessary in preparation for Hippo going public (and scaling) as well as to improve its Underwriting outcomes through effective risk modeling (using ML and traditional actuarial approaches).

I contributed towards both modeling (DS/ML) and backend / data / ML engineering efforts while working on important Data Science problems in Dynamic Pricing and Underwriting / Revenue Forecasting.

In the spirit of data-centric AI, we focused on building pipelines for data acquisition to acquire new and interesting datasets, feature engineering, MLOps, and building the backend to make data science iterations/experiments systematic, reproducible and fast.

At Agilent, I led the development of human-in-the-loop Machine Learning (Collaborative Filtering) powered Recommendation Engine for the digital business (homepage and search) that resulted in several fold increases in different engagement and revenue metrics. We used Spark for large scale data analysis and to build feature engineering and ML training pipelines and also ran experiments to improve the relevance of recommendations.