I have considerable experience with:
|January 2019 – present||Software Engineer|
|2017 – 2018||Engineering Manager|
|2015 – 2017||Software Engineer|
Founding member and Tech Lead for Lead Ads, a News Feed ad product that gives people a quick and privacy-safe way to sign up to receive info from businesses; now responsible for nontrivial fraction of overall FB revenue (cannot share exact numbers).
As TL, I built save/prefill: (improves submission rates by 50%), lead download via CSV (covers 50% of revenue); identified and solved the hardest problems (1% of lead submissions failing); architected and led 2 eng in building new UI that lets advertisers integrate with CRMs right from within Facebook Ads Manager (improved CRM adoption rate by 20%); identified proxy metrics for lead quality (challenge: no ground truth), and drove 20% improvement on those metrics via iterating on form filling user experience: now adopted by 10% of revenue.
As Engineering Manager, I supported 5 engineers; drove major revenue growth by enabling advertisers to use lead forms within other ad objectives (10% incremental revenue in one half); initiated effort to combine client native post-click experiences; provided direction for improving lead quality by exploiting Facebook Pixel as a source for downstream performance.
|2012 – 2015||Software Architect|
Designed and implemented a spend control system for real-time bidding (RTB) campaigns that currently handles over 1 million requests a second across 4 geographical regions and manages over $20 million dollars of client spend.
Conceived, designed and built a real-time audience size estimator that employs hyperloglog estimates, but also adds the ability to perform (a limited range of) logical expressions on audience segments (e.g., “how many people added items to cart, but did not purchase anything?”).
|2010 – 2012||Sr. Software Engineer|
|2007 – 2010||Core Platform Engineer|
Designed, implemented and maintained a significant portion of backend infrastructure for Efficient Frontier’s online marketing platform, including syncing massive amounts of clients’ marketing data into our sharded PostgreSQL instances, a distributed dependency-based workflow engine, early versions of our retail offering, automatic version management for PostgreSQL tables, change history, slow database query optimization (30X speedups), etc.
Gained deep understanding of online marketing: search, social and display (including RTB).
|2005 – 2007||Software Engineer|
Involved in design, development and maintenance of Teragram Direct Answers – a system to provide precise answers to user queries (e.g., “who shot abraham lincoln?”).
Designed and developed a highly customizable “pipeline” for automated, continuous indexing of huge amounts of documents.
Customized both applications into a solution for several major web portals and news organizations, serving thousands of user search queries every hour.
Python, PostgreSQL, Bash, C, Java,
and visualization (
|2005||Sc.M. Computer Science||Brown University, Providence RI, USA.|
|2003||B.E. (Hons.) Computer Science||Birla Institute of Technology and Science, Pilani, India.|
M. Renieris, S. Ramaprasad and S. Reiss. Arithmetic Program Paths. IEEE Foundations of Software Engineering, 2005.
A. Bertossi, C. Pinotti, S. Ramaprasad, M. Shashanka, and R. Rizzi. Optimal Multi-channel Data Allocation With Flat Broadcast Per Channel. IEEE Transactions on Computers, 2005.
Available upon request.