Subham Kumar R Email & Phone Number

Chennai, Tn, In

Chennai, Tamil Nadu, India

Mitigating bad inventory with Ad-IQ analytics• Performed analysis on ad inventory quality to detect bad inventory. Remnantinventory in digital advertisement refers to bad inventory where publishers are leftwith unsold ad slots on the web page during the ad exchange real-time bidding• Modeled extensive lineage by aggregating ad density to quantify ad served capacityacross Condé 32 markets and 21+ brands• Engineered pattern match algorithm to identify and standardize multitudinousvalues of synonymous entries leveraging fuzzy library via python udf in Databricks• Alleviated 20% of bad inventory which helped the downstream teams cater to betterpricing management and cut debugging time by 40%Performance pruning on digital advertisement platform• Ideated and developed a new strategy to integrate and prune the performance of adcampaign delivery across the globe• Led architecture design discussions and implemented API ingestion that enablesauto-healing and is optimized forfast query reads on ∼250m data records per day• Defined extensive context-based data audits to decipher and track down dataanomalies way ahead of failures• Reduced the compute costs by 60% which led to savings of >$150k annually

Chennai, Tamil Nadu, India

▪ Product performance analytics• Collaborated and outlined a data warehouse architecture on Azure data lake to meet growing business needs on product analytics mart leveraging big data tables of size >5+PB weekly• Designed Rest APIs data ingestion ETL pipeline using Azure Databricks to parse JSON/ XML files that included confidential and PII data• Adapted incremental processing framework using watermark implementation on ADF and streamlined a strategy to handle slowly changing dimensions (SCD type2) using data flows, offloading work from SQL Managed Instance (SQL-MI)• Boosted query performance by 25% and enabled daily savings of ~2 hours by improvising major bottlenecks▪ Cloud SSIS migration• Translated MS SSIS legacy jobs to fit into Azure cloud equivalent by deep analysis of the system and segregating dataset into two (Data Factory/ Databricks)• Migrated 500 TB data warehouse including stored procedures, SAP replicated tables, and views from legacy to Azure cloud• Improvised performance by 1.2x using optimization techniques like using higher-order functions to eliminate records from multiple nesting levels, and map-side joins for better cluster performance• Automated advanced orchestration and customized pipeline monitoring job using Azure Logic Apps, turning down resolution time for the QA team by 3x

Chennai Area, India

▪ Data Modeling - Financial ReportTech Stack: SAP BW/HANA, Microsoft Analysis for Office, SAP Analytics Cloud- Built custom HANA views that incorporated flattening of large hierarchical structure representing a binary tree type- It helped to boost query performance by 50% and eliminate multiple joins that were redundant- Scaled the custom view to deliver TB's of querying within a fraction of mins - Integrated the new logic in SAP Analytics Cloud (SAC) for dashboarding requirements that yielded improved refresh rates of 30%▪ Multi-source Ingestion Pipeline - Cost Center Action ETLTech Stack: SAP BODS, Flatfiles, SAP HANA, SAP BW, Oracle SQL, SAP Lumira- Framed generic and optimized ETL solution to fetch data from Oracle/ flatFiles and move it to SAP HANA- Delivered complex query transform logic that addressed data anomalies and resulted to classify datasets (high/ medium/ low)- Took complete ownership and worked closely with business teams to understand the requirements and deliver enriching dashboards using SAP Lumira ▪ Job Approval Automation Tech Stack: Power Automate (Microsoft Flows), Power Apps, Power BI, Sharepoint Online- Architected a simple workflow job that takes care of approval from multiple teams, synchronizes and collaborates to send the team push notifications/ email reminders - Crafted a PowerBI dashboard with metrics like job health, turn-around time, status, and pending/ queued requests which led to ease of tracking and increased turn-around time to 4x- Formulated data ingestion through Sharepoint Online using O365 authentication, eliminating the use of third-party tools that led to cost savings of ~$1750/month

Chennai, Tamil Nadu, India

- Worked closely with the analysis & development team to unveil core performance metrics- Migrating legacy reports from SAP Lumira to Tableau and adding a fresh look to the client's portfolio by crafting a sales dashboard - Active blogger and the first to embark on a new series - R visualization

Chennai Area, India

- Worked with ML algorithms like regression with recursive partitioning using R (script)- Developed an algorithm in the field of Grey Theory for time-series prediction that provides maximum efficiency- Case studies of Delta Air Lines which included predicting the total number of passengers and their revenue for the fourth quarter 2017- Published a regular research paper at IEEE conference under Computational Science and Computational Intelligence (CSCI'2017) titled 'Machine Learning Grey Model for Prediction' administered by American Council on Science and Education- Worked on one of the millennium problems 'The Travelling Salesman Problem' (TSP)

Gurgaon, India

- Communicated and conducted campaigns that promote the thought of an internship- Hailed across multiple boards on campus to bring out the hidden talents - Expanded my professional network working as Student Partner (ISP 7.0)

Bachelor'S Degree, Computer Science, 8.1

Activities and Societies: National level presenter at symposiums | Internshala Student Partner (ISP 7.0) | Volunteered in organizing.

Ssc, Mathematics And Computer Science, 89.4

Hsc, 90.8