- Building horizontal services, systems and datasets supporting multiple marketing channels like SEO/SEM, Display, Paid Social, Email/Push & Affiliates- Attribution framework, campaign management toolsets across platforms and partners- Charting the path for next versions of marketing technology in a privacy focused world, think no 3rd party cookies, Apple ATT etc

System Architecture Improvements• Played important role in designing and improving architecture of multiple internal systems.• Helped in improving the architecture of internal bidding system from single process to distributed system capable of scaling horizontally based on load.• Helped in designing a more dynamic application configuration management system for airflow and potentially other applications in the future.Shopping Campaigns Pipeline (Airflow, Python, Hive)•… Show more System Architecture Improvements• Played important role in designing and improving architecture of multiple internal systems.• Helped in improving the architecture of internal bidding system from single process to distributed system capable of scaling horizontally based on load.• Helped in designing a more dynamic application configuration management system for airflow and potentially other applications in the future.Shopping Campaigns Pipeline (Airflow, Python, Hive)• Refactored existing perl based pipeline to publish shopping campaigns to different ad engines into more robust airflow based pipeline.• Vastly improved the system reliability and trackability of the deploy process.• Achieved over 10x improvement in deployment time over existing system with the ability to deploy millions of shopping campaigns if required on a single day. Show less

Product Back Bone (Scala, Cassandra, Kafka, Spark, Google Protobuf)• Developed 'Store' Dao layer capable of creating/accessing Cassandra or in memory store based on the environment.• Developed configurable cassandra store capable of managing sessions to cassandra with the ability to read/write streaming data from kafka/spark. • Designed/Modeled Cassandra keyspaces, tables and data center configurations based on query patterns and added ability to create/manage the keyspaces from the… Show more Product Back Bone (Scala, Cassandra, Kafka, Spark, Google Protobuf)• Developed 'Store' Dao layer capable of creating/accessing Cassandra or in memory store based on the environment.• Developed configurable cassandra store capable of managing sessions to cassandra with the ability to read/write streaming data from kafka/spark. • Designed/Modeled Cassandra keyspaces, tables and data center configurations based on query patterns and added ability to create/manage the keyspaces from the store.• Used google protobuf for standardizing data schema, enable easy mapping between different domain objects and for platform agnostic serialization of data.Campaign Management Service (Java EE, Spring, Solr/ElasticSearch, Tomcat)• Developed multiple services to bulk publish campaigns/adgroups/keywords/ads to adwords/adcenter which handled extensive validations and data transformation to convert user input into format understood by ad engines. • Migrated data from smaller Solr cluster to much larger ElasticSearch cluster and optimized the indexes improving query and indexing performance by around 15x with improved stability. Rewrote the DAO layer to talk to ElasticSearch. Designed ElasticSearch schema and indexes for our dataset.Data pipelines (Python, Airflow, Apache Spark, Hive, Hadoop, HDFS, ElasticSearch, Mysql)• Developed, tested and maintained multiple ETL pipelines to generate and transform bigdata sets into different formats for different users like data scientists and marketing managers.• Involved in migrating multiple existing ETL pipelines from our homegrown ETL framework to a more advanced and robust Airbnb Airflow framework. • Improved the speed of pipelines by migrating from hive/hadoop based queries/jobs to more faster spark queries/jobs. Show less

SEM Dashboard(Ruby On Rails, Memcached, MySQL, Backbone.js, Coffeescript, Haml, Passenger) • Architected and developed internal dashboard used for reporting analytics and campaign management operations to manage the SEM program. Built the application from scratch. • Sole developer and managed all aspects of the dashboard across the development stack including feature planning, development, schema management, testing and automated deployment using jenkins. • Interacted with other… Show more SEM Dashboard(Ruby On Rails, Memcached, MySQL, Backbone.js, Coffeescript, Haml, Passenger) • Architected and developed internal dashboard used for reporting analytics and campaign management operations to manage the SEM program. Built the application from scratch. • Sole developer and managed all aspects of the dashboard across the development stack including feature planning, development, schema management, testing and automated deployment using jenkins. • Interacted with other REST based services for querying data and submitting operations. • Developed a generic query builder in the UI to build customizable queries with AND/OR clauses and operators like (equal/not equal/contains all/contains any/greater than/less than) with different options displayed based on data type being queries. • Abstracted out the common functionalities of the dashboard like charting/campaign management services/common summaries into Rails engine which could be imported into any rails app. This was done to enable multi-tenant support for different internal SEM customers like Walmart.com, Sams club etc. Show less

HealthRecruit – Search (Ruby On Rails, SenseiDB, PostgreSQL, Backbone.js, Coffeescript, Git, Github, Rubymine) • Built an enterprise search tool for recruiters to search for healthcare professionals imported and classified from facebook. • Used SenseiDB for maintaining indexes and documents for highly performant and scalable text searches of millions of users accessible from facebook. • Implemented caching policies for making searches faster. • Implemented search filters… Show more HealthRecruit – Search (Ruby On Rails, SenseiDB, PostgreSQL, Backbone.js, Coffeescript, Git, Github, Rubymine) • Built an enterprise search tool for recruiters to search for healthcare professionals imported and classified from facebook. • Used SenseiDB for maintaining indexes and documents for highly performant and scalable text searches of millions of users accessible from facebook. • Implemented caching policies for making searches faster. • Implemented search filters based on keywords, location, companies, universities, majors, degrees. • Developed faceted and boolean search logic.HealthRecruit – LeadGen (Ruby On Rails, SenseiDB, PostgreSQL, Backbone.js, Coffeescript, Git, Github, Rubymine) • Designed and built a recommendation system that would surface relevant candidates to the recruiters based on their searching patterns. • Built features that would help recruiters organise and save searches for specific positions they are trying to recruit for. • Implemented foldering and clipboarding features for organising candidates.HealthRecruit – Messaging (Ruby On Rails, SenseiDB, PostgreSQL, Backbone.js, Coffeescript, Git, Github, Rubymine) • Built a messaging system to enable recruiters to contact potential candidates and in-app conversations between recruiters and candidates. • Built messaging channels which included emails and in cases where emails are unavailable, ability to quickly contact a candidate over facebook utilising facebook graph and friend connections. Show less

Graph System – Old Site(Java EE, JSP, C++, Apache Struts, IntelliJ IDEA, Perforce) • Managed the existing friend graph system which stored connections between over a 50 million users. • Resolved issues with the graph system which was partitioned across multiple servers and used MapReduce for executing distributed friend connections related queries.Commerce System – Old Site (Java EE, JSP, Apache Struts, IntelliJ IDEA, Perforce) • Managed and resolved issues with the… Show more Graph System – Old Site(Java EE, JSP, C++, Apache Struts, IntelliJ IDEA, Perforce) • Managed the existing friend graph system which stored connections between over a 50 million users. • Resolved issues with the graph system which was partitioned across multiple servers and used MapReduce for executing distributed friend connections related queries.Commerce System – Old Site (Java EE, JSP, Apache Struts, IntelliJ IDEA, Perforce) • Managed and resolved issues with the existing commerce system which uses ad-commerce hybridization to generate revenue.Data Provider System – Social Gaming Platform(C#, ASP.NET, IIS, MSSQl Server 2008 R2, Visual Studio 2010, TFS) • Developed crucial infrastructure piece which performed CRUD operations using local cache and sql server. • Supported asynchronous updates and deletes using Tasks and wrote stored procedures which performed table creation, updates and get or create operations making a single database trip for enhanced efficiency.SocioPay System – Social Gaming Platform(C#, ASP.NET, IIS, MS SQl Server 2008 R2, Visual Studio 2010, TFS) • Developed commerce system for the gaming platform which involved complex rules pertaining to ad-commerce hybridization and multiple payment providers. • Integrated the commerce system with the external game developers which involved backend callbacks. • Integrated PayPal with commerce system and supported Vindicia for credit card processing.Jobs And Admin Pages – Social Gaming Platform(C#, ASP.NET, IIS, MS SQl Server 2008 R2, Visual Studio 2010, TFS) • Developed multithreaded and asynchronous Jobs framework which supported recurring jobs with failure handling. • Developed Admin page for adding and updating Games to the gaming platform. Show less