Summary
A graduate from IIIT Hyderabad, my first year was with Infosys, after which I co-founded a technology startup Spinaxys Solutions where I worked in products ranging from SMS based personal profiling systems to bluetooth based proximity marketing solution. Going ahead, I joined Synovel, where I designed and lead the development of the open-source Collabsuite server. Then I moved on to Vizury, a startup in the advertising domain, where I lead the RTB platform pipeline and helped build scalable systems by leveraging cloud computing. Moving on, I joined Khoros (formerly Lithium), where over the course of 4.5 years, I've helped build teams, lead teams, set up a lot of the initial processes and culture and help build products. My latest stint was at Meta, where I've got the opportunity to hone my product skills through experimentation and working with XFN partners. Areas of expertise include:
- Deep knowledge in Elasticsearch, including data-modeling and optimizing clusters
- Designing Horizontally scalable systems (my earlier product handled about 2,00,000 reqs/sec)
- Building extremely low latency applications (avg. response time < 10 ms)
- Multi-layered distributed caching systems
- Cloud Computing
- SOA & Restful APIs
- AWS/Network Security (VPN, Firewall, NACLs & Security Groups, VPC, Role-based credentials)
- Continuous Integration Systems (ex. Jenkins)
- Packaging & Release Management
- Monitoring & Alerting
Experience
| Senior Software Engineer |
Mar 2023 - Mar 2025 |
| Meta |
(2 years 1 month) |
- Was part of the Shops Ads measurement team and lead for the first-party measurement pod
- Worked on multiple projects to show better insights to advertisers related to Shops Ads performance. For instance, implemented an onboarding tour in Ads Manager for advertisers who newly started using Shops Ads. This had a high adoption of 15% when compared to other similar features which only had adoption of 7%.
- Work with product manager and design in helping scope out the project and prioritize features that we want to build
- Work with Data Scientist in experiment analysis and evaluate the impact of rolled out features
|
| Senior Software Engineer |
Feb 2021 - Feb 2023 |
| Meta |
(2 years 1 month) |
- Helped build ads for targeting offers in Shops. Shops is a new product that makes it easy for businesses to set up a single online store for customers to access on both Facebook and Instagram
- Logical analysis for identifying product gaps and addressing them to improve product adoption
- Help run and analyze recruited tests for testing the impact of our features prior to production rollout
- Mentoring of team members to help them grow to the next level. Got recognized as a top mentor within the sub-pillar for the same.
- One of the top contributors to the wiki in the team and in the top 5% in the org
|
| Principal Software Engineer |
Jul 2019 - Sep 2020 |
| Khoros India Pvt. Ltd. (formerly Lithium Technologies Pvt. Ltd.) |
(1 year 3 months) |
- Evaluated different approaches for pulling search out as a micro-service
- Was part of a small team to design and bootstrap the project for ES 1.7 to 7.x upgrade
- Helped optimize ES clusters, thereby reducing cluster costs by 50%
- Architecture & Design reviews of features coming out of the Blr office
|
| Principal Software Engineer |
March 2018 - June 2019 |
| Lithium Technologies Pvt. Ltd. |
(1 year 4 months) |
- Lead for the Search team, helping with onboarding, mentoring and implementing product features
- Lead the implementation of Promoted Search feature, which is now actively used by 100+ communities
- Participated in Vista Hackathon and received a special recognition award for an idea based on blockchain to drive user-engagement
- Initiated lunch-n-learn sessions in the Blr office
- Kick-started and mentored the community custodians initiative to help grow knowledge within the Blr senior engineers
|
| Lead Software Engineer |
April 2016 - Feb 2018 |
| Lithium Technologies Pvt. Ltd. |
(1 years 11 months) |
- Helped recruit a lot of the initial talent
- Setup processes for agile software development and on-call
- Prototyped a performance testing framework that was built using micro-services based architecture
- Lead the accessibility team for achieving better WCAG 2.0 AA compliance
- Implemented Virus Scanner Service, which could horizontally scale to handle incremental load as it is gradually integrated with 100% customers
|
| Consultant |
January 2006 - March 2006 |
| Self Employed |
(3 months) |
| Principal Software Engineer |
April 2013 - October 2015 |
| Vizury Interactive Solutions Pvt. Ltd. |
(2 years 6 months) |
- Primary activities include scaling, reducing costs, and ensuring stability and security of the RTB platform, which on an average handles 200,000 requests / sec across six different geographies
- Awarded TECHSTAR for an outstanding performance during the JFM 2015 quarter
- Driving RTB integration with FBX (thereby boosting company revenue immediately by 10%)
- Migration of cache servers from Memcached to Aerospike, a NoSQL database, which has helped reduce cache costs by 50%
- Mentoring migration of our RTB stack to make it more secure by using VPC, NACLs, Security Groups and Roles
- Explore, deploy and configure a monitoring framework for handling dynamic / adaptive thresholds using bischeck, a Nagios plugin
- Architect and implement real-time logging for processing about 200,000 log events / second that are generated by RTB bid servers
- Design a pre-production testing framework using TestNG which helps to identify bugs just before production release to greatly reduce the chances of things breaking in production
|
| Member of Technical Staff |
January 2013 - April 2013 |
| Lexity |
(4 months) |
| Co-founder |
April 2006 - May 2008 |
| M/s. Spinaxys Solutions |
(2 years 2 months) |
| Software Engineer |
June 2004 - September 2005 |
| Infosys |
(1 year 4 months) |
Significant / Interesting Products & Projects
| Bootstrapping ES 1.7 to 7.x Upgrade |
| The ES 1.7 to 7.x was a bottom-up proposal to upgrade elasticsearch for the community platform. Bootstrapping this project was pretty complicated due to the fact that elasticsearch powered a lot of features in the platform and a though evaluation needed to be done for the right approach to be taken, so that we could eventually transition to the final goal of pulling search out as a service by making the right tradeoffs between speed of development, keeping infra costs low and showing early wins. The design was presented in one of the ES meetups: https://www.youtube.com/watch?v=Bkf5w2x1ACs |
| Real-time Logging for Bid servers |
| Due to some requirement, we needed to make our bid servers stateless and therefore had to implement real-time logging. Under this implementation, log events get pushed to a Kafka broker as soon as they are generated in the bid servers. These log events are consumed continuously and written to files by a cluster of servers and uploaded to S3 for later processing by Analytics. We faced a lot of challenges in the bigger goes where at peak times, we have up to 50 bid servers, with each bid server pushing 4000 log events / second of avg. 500 bytes each. Cumulatively, this comes around to 100 MB of uncompressed log data to be processed every second. Through various optimizations such as compression, using flume, implementing a custom kafka partition rebalancer notifier, and resolving all the disk and network bottlenecks, we were finally able to implement a robust real-time logging architecture which ensured data buffering during broker downtime, no duplication of data and a horizontally scalable cluster of log processing servers. |
| FBX RHS And NewsFeed Integration |
| Integration with FBX had two major challenges - 1. A strict timeline of 4 weeks from FBX within which integration had to be completed and 2. The non-standard way of delivering ads by sending ad information in the bid-response itself. Meeting the deadline required stretching myself and the non-standard way of serving ads implied setting up the new feed infrastructure from end-to-end on the RTB side. FBX Integration turned out to be the major revenue booster for the year. Here are some numbers which show the impact of the integration within a couple of months of it going live - 30%-40% revenue contribution in whichever campaigns it went live, avg. profitability of 55% compared to that of 38% for non-FBX and 45% of the total RTB impressions (40-50 million). |
Skillset
| Interests | : | Programming, System Administration, System Design, Product Architecture, GNU/Linux Debian, NoSQL |
| Languages | : | Java, Python, Hack, React.js, C/C++ (prior experience) |
| Tools | : | Nagios, Elasticsearch, Aerospike, MySQL, Redis, Memcached |
| AWS | : | EC2, RDS, S3, Auto-scaling, ELB, SNS, SQS, IAM, AMI Creation, Cloud Formation |
Education
| International Institute of Information Technology |
2000 - 2004 |
| Bachelor of Technology (B.Tech.), Computer Science |
(4 years) |
Recent comments