Sridhar Ramasamy

Areas of Interest:

Technical Skills

Work Experience:

Sr Software Engineer   Oracle America Inc, Bellevue WA   Sep 2018 - Current

• Project: Advertising Data Pipelines (March 2023 - Current):>
  
  • Handle voluminous Kafka stream data with variety of records and make it available for the org. to consume.
  • The data pipelines receive 1000’s of records every minute and are processed in OCI spark dataflow.
  • The data is made available in both AWS and OCI, while processing happens only in OCI.
  • Migrating our workflow from AWS to OCI, setting up new orchestration of pipeline using airflow/argo.
• Project: Data API/Data Collaboration Platform (July 2021 – March 2023):
  
  • Designed model to facilitate data storage, access, and enforce data lease contracts.
  • Data API defines schema, transformation and joins across schema, source and destination location.
  • Designed API for data model, entities, and its relationships & persistence.
  • Usage of IAM roles/policy to allow limited access to a leased data location and enforce contracts and TTL.
  • Designed and built security model for the Data Lease Service with RSA based auth and auto JWT generation.
  • Integrated with other systems to effectively manage source data and lease data according to contracts.
• Project: User Consent Management System (July 2020 – July 2021):
  
  • Opt-out data auditor will be capable of keeping a register of dataset, export opt-outs and trigger alerts to dataset owners and generate periodic reports and keep counts.
  • Load testing the service using Apache JMeter for an endpoint that needs to capture few million records.
  • Productionizing Apache Pulsar cluster with TLS certs and setting it up on OCI Kubernetes.
• Project: Dynamic Data Packaging and Delivery Project (Sep 2018 – July 2020):
  
  • Designed and implemented a new data packaging & delivery project to replace an existing batch processing.
  • The project saves about $2M/year in computing cost, brought down compute time from 96 hours to 14 hours.
  • The different supporting pieces are - Data selection, Data Matching, Data Transformation and Packaging, Data Encryption, Data Compression and Data Delivery.
  • Automatically deliver reports and notify the stake holders, clients, and dev team about the deliveries.
  • Interacted with the Product team gathering requirement for the system & drafting functional specification. Train the operations team with the new system and oversee on-time resolutions of support tickets.
• Tech Stack: Java/Spring Boot/Kubernetes.
  
• Data Processing: Apache Spark/Qubole/AWS.
  
• Job Orchestration: Luigi/Apache Airflow.
  
• Migration of all our CI/CD pipelines from Jenkins to Gitlab. Replicate all the functionalities of existing pipeline. Perform unit testing and integration testing to seamlessly integrate with other existing systems.
  
• Coordinate with other teams/pods to enhance design/development and support cross-team projects.
  

Software Developer   Red Hen Systems, Fort Collins CO   Nov 2016 - Current

• Led a Field Report creation project for Anadarko on C#/.Net to view and publish survey data.
  
• The paper-based forms were converted to digital application based on a .Net Windows platform capable of data entry, sketching with stylus or touch pad and attaching photos.
  
• The tool is capable of formatting the data in the pdf, resize images/customer sketch, publish data to a database and send an email with the pdf as an attachment.
  
• Responsible for development of IsWhere, Media Geo-Tagger softwares - C#/.Net/WPF/Forms.
  
• Introduced online licensing & management system for all our software products reducing manual effort.
  
• Improved the software by adding tools to clip videos, modified UI, provided multiple file compatibility.
  
• Batch automation process for ETL workflow – Data extraction, processing, analytics, and delivery.
  
• Worked on GIS projects for various oil & natural gas companies, universities, and government agencies.
  
• Develop new features and maintain IsWhere and Media Geo-Tagger software, based on C#/.Net.
  
• Batch file automation of video files copying, decoding, creation of CSVs/XMLs, ArcPy processing (.shp). Reduced time taken from 2 working days to just 4 hours which is a 200x improvement.
  
• Introduced support for Flexible & Interoperable Data Transfer (.fit) files for Garmin VIRB’s.
  
• Optimization of Python codes to process several hundred thousand spatially referenced data points for data analysis of methane PPM, wind vector data to create KMLs.
  
• Maintain ArcGIS server - MS SQL server and the front end client to display GIS results.
  

Grad Student Researcher   Computer Networking Research Lab, CSU   Feb 2015 - may 2017

• Used the principles of low rank matrix completion to capture and reconstruct the topology of undirected graphs.
  
• Topology reconstructed using two different measurements,
  a) Shortest hop distances from nodes to a set of nodes called as anchors
  b) Shortest hop distances between any two random node pairs
  
• Applied on Wireless Sensor Networks and Social Network Datasets (Facebook, Collaboration and Email networks)
  
• Compressed representation of Social Networks
  
• Utilized the extended Robust Principal Component Analysis [Randy et. al.] for predicting the unknown entries of low ranked matrix.
  Wireless Sensor Networks
• Spatial Analytics on Wireless Sensor Networks and GPS free localization using topological coordinates [Jayasumana et. al].
  
• Created new 3D test networks for WSNs. Number of nodes ranges from 500 to 1600.
  
• Reconstructed the topology of WSNs with only 2% to 5% of coordinates from distance matrix
  Social Networks
• Analysis of the social network properties,
  a) The three test networks denote three different clustering
• between 0 and 1 (0.2379 to 0.6048) (Start to Clique)
  b) Very different Avg. path length (2.55 to 6.048) and diameter (5 to 18)
  c) Number of nodes vary between 750 to 4200
  
• Reconstructed the topology of the social networks within 1 hop distance error with only 1% of the coordinates.
  
• Used Python (Pandas, Numpy, Scipy, Matplotlib, Networkx), Gephi for the implementation
  

Education:

Projects

Publications:

Links: