At Google Ads, I re-architected the UMI Massager workflow for Privacy Centric Measurement, replacing a complex monolithic post-processing pipeline with a configurable platform of discrete, debuggable steps. This improved my team’s development velocity and gave other Google Ads subteams a path to onboard their post-processing logic.
I designed reusable components for complex fanout joins and CASE/SWITCH-style rule-based selection, and extended the framework with date-based calculations and multi-level flooring support. I also debugged MapReduce pipelines and tuned performance around serialization and data shuffling.
At Squarepoint Capital, I maintained market data infrastructure and provided Level 2 support for live feeds. I revamped scheduler configuration from per-region to per-feed files and added XML-based controls for muting individual multicast lines.
As a Software Engineer on Squarepoint Capital’s Market Data team, I maintained C++ exchange feed parsers and developed ’tique’ to process direct exchange multicast and TCP lines for quant repositories. Earlier, I built a Python tool at Bajaj Auto Limited to post-process vehicle test data and automate manual tasks.
