Founding Team | Head of Technology | Architect | Engineer

- Individual contributor on Near's DSP offering - Engage and RTB and Near Data Processing Platform- Served as a member of Near's Architecture Board (ARBO), contributing to strategic decisions and technical architecture planning.- Focused on cost reduction and enhancing the efficiency of Real-Time Bidding systems, resulting in increased QPS capacity for existing machines and reduced overall latency.- Successfully scaled the RTB system to handle approximately 200,000 Queries Per Second (QPS), generating over 13 terabytes of data daily, achieved through optimization on a lean cluster.- Managed Bidder integrations with some of the world's largest Ad Exchanges, ensuring seamless and efficient collaboration.- Spearheaded the re-architecture and implementation of a new ad search flow for the RTB, enhancing efficiency and responsiveness.- Owner of Near Places - oversaw integrations with major geo-data providers and led a team of 5 GIS engineers for Near Places.- Conceived and created an Automated Pipeline for refreshing comprehensive GIS data from diverse sources, achieving a significant performance boost (30x50x) during data refresh.- Implemented Deployed CI/CD Pipelines across multiple projects within the organisation.- Established an R&D pipeline enabling users to extract building footprints (Polygon Geojson) from Satellite Imagery and seamlessly integrate them into the Near GIS system.- Collaborated on various systems throughout the organisation, enhancing their performance and scalability.Technical Skills - Java (Springboot / Dropwizard), Python, Elasticsearch, MongoDB, Redis, Kafka, Spark, Amazon Redshift

- Core team member on re-architecture and launch of Engage, Near's self serve DSP.- Played a major role in migrating all of Near's infra from Softlayer bare metal servers to AWS cloud.- Engineered and implemented a real-time entity recognition solution in Golang, achieving response times of less than 5 milliseconds.- Architected, owned, and maintained PlaceMatrix, a pioneering solution in the digital advertising space.- Effectively managed an off-site vendor team focused on geo data content creation. Enriched brand coverage of Points of Interest (POI) content, optimizing advertising expenditures for customers.- Handled integrations with various data partners, including Master Card, Experian, HERE Maps, to enhance Near’s niche location focus in advertising.- Led a 2-member team on Near's Real Time Bidder system, ensuring its efficiency and seamless operation.- Led a 5-member team (1 UI + 2 backend + 2 data) on CARBON, Near's ID-level Data Platform, overseeing its development and performance.- Built the entire Data Processing Platform for CARBON- Developed Near's Pixel ManagerTechnical Skills - Java (Springboot / Dropwizard), Python, Elasticsearch, MongoDB, Redis, Kyoto Tycoon, Clickhouse, Twemproxy, Kafka, Spark

- Core team member on Allspark, Near’s Audience Cloud.- Re-architected and re-built Near's RTB using Java Dropwizard framework.- Architected and developed ID Manager, Near’s central repository for all device, cookie, and user IDs.- Designed, built, and maintained the entire system, featuring a Kyoto Tycoon cluster spanning 3 geographies, managing close to 2.5 billion IDs and growing at a rate exceeding 4 million per day.- Orchestrated a system that efficiently handled approximately 4 billion requests daily, peaking at a QPS (Queries Per Second) of 85,000, maintaining a mean latency of less than 5 milliseconds.- Implemented Geohash for real-time location refinement and correction using Georaptor. The micro-service delivered responses in microseconds.Technical Skills - Java (Dropwizard), Python, Golang, Elasticsearch, MongoDB, Redis, Kyoto Tycoon, Pika, Twemproxy, Kafka, QGIS

- Spearheaded the development, management, and maintenance of Near Places, Near's Geo Data Repository, featuring a highly scalable and low-latency REST API system.- Architected and constructed the entire cluster that powers Near Places, ensuring seamless operation and optimal performance.- Conducted thorough evaluations of project requirements and provided technology recommendations to support the construction, maintenance, and performance of the system.- Established a comprehensive database comprising more than 75 million places, meticulously building polygons and geographic region polygons sourced from multiple channels.Technical Skills - PHP, Java, Python, Lua, Elasticsearch, MongoDB, Redis, Twemproxy, Kafka

- Integral member of a 2-member team responsible for managing and maintaining Near's Real-Time Bidding (RTB) system.- Successfully scaled the bidder from 5,000 Queries Per Second (QPS) to 25,000 QPS, enhancing the system's capacity and performance.- Implemented crucial changes to the RTB system to facilitate the inclusion of Video and Native ads, expanding the platform's capabilities.- Led the migration task of transferring Near's Unique ID data to a Redis-based store, optimizing data storage and retrieval processes.Technical Skills - PHP, Lua, Elasticsearch, Redis, Twemproxy, Kafka

- Loog - HAProxy Log Analyzer: Lua-based UDP server to read HAProxy logs in real-time, process, and provide analytics based on the same (table and graph). Handled about 50K QPS per instance on the Near Real Time Bidder. This used Redis as the data store.- ReqLoc: PHP-based tool using Elasticsearch as the primary datastore to curate and fetch total as well as aggregated counts based on user IDs, Operating Systems, and Carriers.- RTB Test Tool: Tool for a pre-deployment check of the Real Time Bidding System.Technical Skills - PHP, Lua, Elasticsearch

Organizing and voluteering all the Training and Placement related activities in the college. Interact and communicate with various HR managers from different firms that visit the college.

Learned the development, implementation and maintenance of the payroll management system being used in the firm.