Privacy (2021 - )* "Cofound" Regulatory Readiness pillar in Central Privacy to scale x-Meta regulatory initiatives with "unprecedented" progress (VP's shoutout)* Grow a team of 15 to an org of ~50 (incl. 5 M1/2s; 90+ with XFN) in 1.5 years. Recruit & grow 8 IC6+. Supported M1->M2, IC7->IC8. Articulated & align the archetype to unlock progress * Scale Consent UX Platform and reduce dev cost of GDPR-style flow by ~18x (110 eng-months -> 6)* Drive new commitment-grade solutions, with v-team (~10 orgs, >80 eng), to systematically track consent-governed data in online & offline stack. Enable field-level granularity policy control at scale (1st time at Meta). Play pivotal role in new relationship building, influencing org's roadmaps without authority, (re)alignment of resourcing, and supporting v-team through pivots and unforeseen disruptions (e.g. major reorgs)Ads (2018-2021)* Supported an org of ~55 (incl. 5 M2/TLM2/M1s; 85+ with XFN). Recruited M1/M2s and landed IC5->IC6->IC7 growth cases* Developed x-Meta strategies, with XFN, which generated $billion incremental revenue (2020) in Ads guidance and creative areas. Note: strategies still in play 2022

Built a 0->1 product from start to general availability and revenue generation in <1yDid whatever it took to get this done: recruiting, building a team, people management, architecture, hands-on coding. Mix of scala and python. Used Finatra, Kafka, MongoDB, HDFS, ElasticSearch, Docker etc.

* Explored to start my own. Didn't get traction, humbling experience* Advised startups * Worked with deep learning startups: a) MetaMind - acquired by Salesforce (05/15-08/15) b) Pilot.AI (01/16-04/16, angel investor)

* Led API perf efforts. Enabled Uber, for the 1st time, to sustain historical load from New Year's eve traffic with no incidents. Supported the teams owning all business logics: API, payments, comms, internal tooling for city ops, driver/rider safety and experiences

Led the innovation and development of data-driven products (Trends, Events, Discover, WhoToFollow, @MagicRecs, @EventParrot.) with ~60 folks (incl. 6EMs; across SF/NYC) spanning infrastructure, data science, web, iOS and android Highlights:* Grew the org from ~20 to ~60. Championed and got two initiatives/teams funded; Grew EM bench with 5 IC->EM and 1 external hire* Pitched and landed acquisition and integration of Spindle, who played critical role in building out Twitter's TV/sports experience * Initiated and deployed two high leverage ML infra (Trending X in Y, Engagement Predictor)* Led the 2014 #WorldCup efforts with algorithmically driven timelines, across 12 languages/markets with over 50% user return rate for the event* Initiated ideas and supported team efforts to innovate with real time recommendations and breaking news detection (e.g. EventParrot)Note: 10/2012 - 06/2013: Senior Engineering Manager

* Shipped 1st version of LinkedIn influencer* Led the acquisition/integration of Thoora and bootstrapped Content Relevance area to accelerate LinkedIn's content strategy* Support the teams responsible for LinkedIn Today, Groups and Influencer

Led Search/Typeahead team through 5m->150m hyper-growth. Responsible for search products, quality and infra across LinkedIn's top use cases (People/Jobs/Companies/Groups)Highlights:* Grew the team from 5 to 15 and supported 3 IC5->6 cases* Scaled out the people search system to handle member hyper-growth (5m -> 150m)* Increased engagement by 225% YOY via faceted People search efforts. Revamped LinkedIn's Company and Group search products * Landed scalable 2nd-degree connection typeahead feature with 20ms average response time (open sourced: Cleo). Note: lacked viable solution for years * Reduced search page load time from 3.5s to 1s by optimizing across stack, especially on front end. For example, early flush cut ~100-300ms across all LinkedIn pages* Reduced indexing time from weeks to hours and query serving latency by ~200ms via redesigned Search Content k-v Store

* Designed and implemented a query understanding system which internally performs "advanced searches". A critical search quality component with strict latency budget (<5ms, called for each query). Implemented with NLP algorithms, segmentation, probabilistic field tagging and query expansion.* Built and maintained search indexing pipeline* Implemented and owned LinkedIn's widely-deployed memcached client (Java) with consistent hashing, UDP transport support, automatic partition failure detection and recovery mechanism.* Hands on expertise in Lucene ranging from analyzer, parser, to index structures and optimizations.