Exploring Glean's mission of "Work AI for all" to make all workplace knowledge accessible and actionable.

* Realtime Voice Assistant

* PyAgents infrastructure overhaul

* "Glean where you are" browser extension

* few more projects

Pursued different missions to help underprivileged and marginalized groups through job search, education, financial wellness, and social justice. Navigated stakeholders, policy and legal challenges, and external partners, creating space for teams (2 to 5 directs, total 5 to 15) to build effective and compassionate products. See my resume in profile > contact info > website for Google blog posts.

Projects:

• Big Moments (2020 - 2023): Led a team to create Search features to provide deeper context for news stories on Google such as the Ukraine war and the Oscars and other sensitive/popular moments such as protests, celebrity deaths, mass shootings, and media/sport events. We show “lives affected” from Reuters for the scale of a war or mass shooting, live images from Getty for an on-the-ground view, in-depth news explainers for the Roe v Wade overturn and others, “How to help” to donate to the Ukraine refugee crisis, human voices from social media for context on incidents like the Will Smith slap, podcasts, and award nominees and winners for the Oscars.

• Financial Wellness (2020): Advised teams to help with unemployment benefits especially for job seekers during the pandemic. Created features that improved financial literacy.

• College Search (2017 - 2019): Led a team to launch Search features for college name and list queries, to help underprivileged students make informed college choices by 1) avoiding predatory schools, 2) recognizing lower-cost options, 3) understanding ROI, 4) choosing programs aligned with their career goals, 5) seeing student life and if they will fit in, and 6) getting notified about financial aid.

• Job Search (2015 - 2017): Led a team to launch a job search feature for queries like “accountant jobs near me”, by collaborating with major job boards and aggregators to enable the user to search for and filter jobs to identify relevant ones, understand estimated salaries and benefits, look for remote work, and be notified of new jobs.

Led search infrastructure projects such as web and knowledge graph indexing, building large, fast, and complex systems directly in data centers before the advent of Cloud technologies.

Projects:

• Foundation (2014 - 2015): Worked on unifying all of Google’s Search infrastructure including private/public, structured/unstructured, crawl/indexing/serving systems into a single “database of everything” and shared business logic with modular APIs and microservices.

• Livegraph (2012 - 2013): Built the incremental indexer for the Knowledge Graph, with low-latency reconciliation and composition. This powers Knowledge Panels and many other Search features.

• Alexandria (aka Caffeine) (2007 - 2013): Led the team to build the incremental indexer for Web Search that scaled to 1T web pages, averaged 1 min latency, and reduced staleness by 50%. It was built on a 250 PB Percolator based transactional data store running on 15,000+ machines, and had innovative techniques to handle diverse content types, redirects, duplicates, hyperlinks, pagerank, webmaster errors, and hacked sites.

• Indexing pipeline (2005 - 2010): Built a batch indexing pipeline that scaled to 200B web pages with 36-hour latency, implementing a virtual segmented 10 PB repository that was needed to ingest a continuous crawl.

• Index scaling (2003 - 2005): Scaled the index from 350M to 8B web pages, doubling index size over 3 weeks on the eve of Microsoft’s 5B index launch (Press: "Microsoft launches the world's biggest index, second only to Google").

Publication: Squirrel: A decentralized peer-to-peer web cache. Sitaram Iyer, Ant Rowstron, Peter

Druschel. Published in the Symposium on the Principles of Distributed Computing (PODC), July

2002, Monterey, CA.