Palash Kulshreshtha
Contact
67 shanti sarovar
Ramghat Road
Aligarh
trickypalash@gmail.com
+918586008714
Work experience
paytm Jan 2020 — Present
- Sharded paytm's central customer side order database to remove single point of failure and increase throughput. (team of 5)
- Active participant in root cause analysis subteam which works on major outage analysis.
-
Responsible for designing and upgrading existing rabbitmq pipeline for order status updates to kafka for durable performance over peak loads and increased consumers.
- Designed in team of 4 central api gateway with real time config reloading, rate limiting, quota limiting, authenticated proxy.
- Managing 3 core elasticsearch clusters serving platform critical usecases.
- Designed percentage wise deployment system where we route few percent of traffic on new build before further rollout. It prevents full traffic impact in case of any issues during deployment.
paytm June 2015 — Jan 2020
- Solved the challenge of uneven slave loads during peak time by rolling out slaves behind proxysql architecture, achieving 99% evenly distributed read load.
- Responsible for providing merchant side order search capability by realtime data indexing in elasticsearch via logstash rabbitmq pipeline.
- Designed and engineered the optimization strategy of a service, by utilizing RPC over rabbitmq, attaining a speedup of upto 800% .
- Designed custom metrics to better monitor overall service health, ensuring higher mean time to failure.
-
Scaled system from batch oriented revenue calculation framework to real time multi step calculation data pipelines.
- Scaled system from few reports to 20,000 reports daily.
Education
B.Tech,Z.H. College of Engg. & Tech., 9.9 CPI 2011 — 2015
Sr. Sec. School Certificate, A.M.U. SSC Boys, 87% 2010 — 2011
Hr. Sec. School Certificate, O.L.F. CBSE, 95% 2008 — 2009
Major Technologies / Languages
- Currently working on nodejs, mysql, redis, elasticsearch, rabbitmq, kafka, proxysql, logstash
- Building upon following aws resources ec2, load balancers, auto scaling groups, S3, lambda, ebs, cloudwatch
- Personal programming experience in C,C++,python, R, Docker/docker-compose from personal projects
- Understanding of Hadoop framework: Filesystem & Mapreduce
- Frameworks- Casperjs, phantomjs, mocha, puppeteer
Projects & Summer Schools
Cli tools for jira and bitbucket
Created command line utility with over 200 Stars for jira and bitbucket which saves at least 30 second per issue or pull request creation
Data scraping for ngo working on water quality data using casperjs & phantomjs [phantomjs/nodejs]
NGO's working towards analyzing and working on water quality projects require consolidated data. This problem of data inaccessibility was solved with web automation using phantomjs and casperjs
UDF for mysql [c++/mysql, make]
Solved the frequent problem I faced while querying my personal databases which was the absence of aggregate function which can give the value of a column, corresponding to which another column is maximum in an aggregated group.