- Leading and mentoring a group of talented and energetic engineers - Proud to be a part of flagship Bangladesh startup, operating in more than 60+ cities in 2 countries, serving millions of people with more than 100 products over time, with only around 50 engineers!

> Lead a team of brilliant data engineers to build a scalable and manageable data platform from scratch at the time when the startup was struggling with high volume data.> Created a data pipeline that could ingest transactions data from 20+ sources and streaming data at unexpected volume (50 GB/day) and velocity (10k/sec). Using Spark, Dataflow, GCS, Airflow> Maintained 10TB BigQuery data warehouse. > Built an allocation system that can match supply and demand at scale.

* I have been working on steaming feature generation on ad, adspace and other combination interaction and view-ability rate on more than 150 GB daily data log using Spark Streaming to boost Recommender System performance.* Developed high performance and low latency micro-service for recommender system which consumes multiple feature and using different algorithm to produce best recommendation within 2-7 milli seconds.* I have been working on enriching Recommender System using different machine learning tools like Xgboost.* Added Deep Learning model using TensorFlow to boost Recommender System performance.* Gained more than 20-50% improvement in ad interaction rate using Recommender System.* Writing script for feature generation and batch recommendation which consumes more than terabytes of data daily. * Migrated etl jobs from cron to Apache Airflow scheduler, managed job dependency easily and handled failing job automatically.* Worked on survey service that runs to measure brand impact on user viewing the ad.* Worked as a scrum master for the small team of data engineers in Dhaka.

* I had been working on cloud based, large scale data mining and data processing architecture (AWS env using Lambda, Sqoop, Hadoop, Hive, Mahout, Hbase ...) to facilitate data driven services.* Started working on Recommender System using collaborative filtering method where aim was to show user less intrusive ads based on users' previous interactions and browsing history.* Exposing number of micro-services to our Widespace's AdServer to provide user profiling and insight data.* Worked on geo service based on PostGIS which was used to position user location to smallest grid address.* Worked on finding similarity between user's based on user different digital footprint and browsing history.

* Worked on high performance low latency Ad Server. * Worked to migrate some of our data loads from RDBMS to Cassandra cluster, running on AWS. * Stared building big data framework using Hadoop in EMR, Pig, Hive etc.* Creating reports on big amount of data using Hive.* Worked on ETL and data pipe line using Sqoop and other tools.

* I had been working here on a rich web editing tool, heavily using javascript.* Enabling writers on the editor to correct his english grammar and asses his writings quality, backed by language services provided by NLP engine. * Implemented large scale document storage system where user can find his document with tag or name arranged in drive.* Developed collaborative writing feature where multiple persons can work on same document at the same time.* Worked on backend services using Java and Spring Framework.

* Leading a small team of mobile app developers.* Design and developed custom UI for mobile applications in J2me, at that time rich UI like now wasn't available that much.* Developed mobile client for e-card (mobile wallet) application.* Developed social networking application mimicking youtube and facebook, but focusing on location based connectivity.