talk-data.com talk-data.com

N

Speaker

Ninad Sohoni

1

talks

DSA Databricks

Ninad Sohoni is a Delivery Solutions Architect (DSA) at Databricks, where he partners with customers to solve complex business challenges using data and AI solutions. With over 15 years of prior experience as a technology consultant, Ninad has led large-scale complex programs to build end-to-end technology solutions enabling organizations achieve strategic and operational goals. Outside of work, Ninad enjoys playing board games, and exploring the scenic trails of the pacific northwest.

Bio from: Data + AI Summit 2025

Filter by Event / Source

Talks & appearances

1 activities · Newest first

Search activities →
Entity Resolution for the Best Outcomes on Your Data

There are many ways to implement entity resolution (ER) system — both using vendor software and open-source libraries that enable DIY Entity Resolution. However, generally we see common challenges with any approach — scalability, bound to a single model architecture, lack of metrics and explainability, and stagnant implementations that do not "learn" with experience. Recent experiments with transformer-based approaches, fast lookups with vector search and Databricks components such as Databricks Apps and Agent Eval provide the foundations for a composable ER system that can get better with time on your data. In this presentation, we include a demo of how to use these components to build a composable ER that has the best outcomes for your data.