DHINESH KUMAR

Sales and Operation Efficiency:

Technology: Palantir foundry, Pyspark, Databricks

• As a lead engineer, worked with product managers to finalize requirements, designed a system to track orders and deliveries, alert users on delays, foresee projected inventory for next 14 days.
• Resolved a key data extraction issue in an upstream Transportation API using recursive solution to track the deliveries better.
• Built an AI Agent that can parse instruction manuals in PDF, help resolve queries based on uploaded documents. It needed vector index to store the documents, convert user query into vector embeddings to perform a search and respond appropriately using LLM.

Goldman sachs (Nov 2019 to Dec 2023)

Elastic search client migration:

Technology : Java, Elastic search, SpringBoot, JUnit, Git

• JEST clients were used to read and write indices from elastic search but are deprecated from ES version 5.6, hence there was a need to provide a robust framework to migrate clients with minimal impact.
• Migrated more than 12 indices which use multiple APIs like write single document, bulk write documents, read single document and read multiple documents.

Risk profile calculation system (FCC):

Technology: Java, Db2, SpringBoot, JUnit, Git

• Worked on design and build of Mortgage risk, IBD risk for potential/existing clients by extracting the information from data lake, categorizing them based on volume/unit of transactions, integrate it into overall risk profile.
  

Conagra (Apr 2023 - Mar 2025)

Technology : Pyspark, Databricks, ADLS, Git, Azure devops

Data Product framework (Medallion architecture)

• As a team lead, designed and built a data engineering framework that ingests, compute incremental feeds, curate and aggregate supply chain data from various sources like SAP, OTM, etc.
• Implemented a reconciliation framework for data across various stages (bronze, silver and gold) to ensure there are no gaps/mismatch across different stages.

Affiliate marketing analytics

Technology: Java, Scala, DynamoDB, SNS/SQS, AWS KCL, DDB Stream.

• Clicks on retail websites are intercepted and redirected to AssociatesTrackingService if it is tagged.
• These clicks are sent to ATP via AMP queue which will eliminate the clicks that are not recordable and publish it to A3.
• A3 uses these clicks to join with impressions and reward the associate responsible for the click.
• Worked on migrating clicks from AMP queue to DDB stream seamlessly (Java).
• Worked on analyzing the clicks received in our service with the clickstream clicks to figure out the missing clicks in our system.

Finacle treasury

Technology: C++, SQL Developer

• Finacle Treasury is an integrated front, middle and back office treasury system that provides full straight - through - processing (STP) and covers the full breadth of instruments across money markets, foreign exchange, fixed income products and their derivatives.
• Worked on Hedge module, which requires intensive calculations by processing millions of records in the database
• Created a separate server for calculations hence making optimum number of server calls to do the calculation simultaneously and efficiently
• Worked on integrating python interpreters in C++ application for market rate prediction using linear regression and dollar offset.

McAfee enterprise anti-virus

Technology: C++, XML

• It is the next generation Anti-Virus which has plug-in play architecture to insert Blades (Firewall, Web Protection, and Threat Prevention) anytime.
• It is used in Enterprises where an admin will maintain the policies for all the clients.
• It is a self-extracting installer that gets installed and after installation, it does content update, product update and policy update
• Developed a patch update framework which would pull patches from the download server and install it in the client machine and report it back to Security Center (web page)
• Worked on fixing Coverity issues, Performance issues and memory leak issues and crashes.

Siebel CRM hospitality

Technology: C++, Siebel Tools on Windows platform

• Implemented a workflow which fetches all the records in the screen and perform various discounts on each record (i.e. Projector, TV, Breakfast) like volume discount, header discount, category discount and write all the records to the Database in a single step and reduce the performance time
• Created Automation scripts using ATF, QTP and worked with code coverage tools.

• Developed SQL procedures using T-SQL to simulate N VMs (Virtual Machines) and reduced the effort and time spent in BAT Testing
• Developed BaseRateCalculator tool in Adobe Flex for CBM (ChargebackManager) which suggests the Server Provider to charge VMs for Memory, processor and HardDisk usage.