Sudhanshu Sharma

• OPT-IN Verification ( VoiceToText NLP Analytics – Live project) - Scope: OPT-IN is consent of customer for marketing via opt-in channels Landline, SMS, Email, and Mobile
• Deliver end-to-end event-driven microservices-based MLOPS pipelines on Kubernetes for OPTIN to cover data preprocessing, resampling, STT transcriber, customer data enrichment in the transcript, Segment Search, SpaCy, Salesforce patch, etc
• Preprocess data Human Transcripts + Voice files for training and test ML models - Custom acoustic NLP Model Training with human-generated transcripts + resampled audio in the German language
• Made continued efforts in data preparation to increase overall accuracy from 0.67 to 0.92 and made it to production which can save ~2 ml € per year of third-party manual verification cost to E.ON
• EDG Data Mesh (On-going) - Scope: Provide raw data, Salesforce customer data, and intermediate data generated by different AI projects as a product to be available for intra-team
• Building concept and design architecture for data mesh
• Making Opt-in transcription available as a data product
• Customer Happiness Index ( VoiceToText Sentimental Analytics - Proof of Concept) - Scope: Customer Happiness Index is a Microsoft Hackathon project prototype which went successful
• Develop end-to-end event-driven microservices on Azure function for MLOPS pipelines which deliver resampling, STT Analytics, Opinion mining, and Sentimental analytics with Cognitive Services
• Deliver UI/UX dashboard for real-time processing demo and Grafana Dashboard for deeper analysis
• Scored winner position in MS Hackathon at Munich and presented future use case to E.ON CXO.

Smart Meters Analytics (SMA) project (Live)
- Scope: E.on Smart meter analytics project focused on analyzing customer’s Smart meters performance metrics and compare them with real-world environment data.

E.ON TargetBI project (Live)
- Scope: E.ON have more than one sister companies. To create a unified MDM (Master Data management) for account, sales, BI data, EON is drawing data from Powercloud Innogy in order to make it available for analytical purposes i.e. Powercloud CDC Data migration from AWS cloud to Azure Cloud with AWS DMS (Database migration service).

LeanIX Integration Pipelines (Live)
- Design and Develop cost effective and secure solution to develop pipelines to integrate LeanIX source with multiple target locations in On-prem, Azure and multiple SaaS platforms like ServiceNow, Coplanner, Symbio, AxonIvy etc

• Autonomous Driving Data Management and KPI (Live) - Design and Develop a Data pipeline for exchange driving captured recordings data between client’s NAS storage and partner’s AWS S3 bucket using IAM Allow role policy with AWS
• Secure Token Service for data security
• Design catalog management of processed data with Hive and HBase
• Design and develop a track management data model and centralized ORM application library for MySql and Oracle
• Cloudera Hadoop Cluster as a backbone for Centralized Data Lake, DAG Workflow management, distributed transformation and data-sharing services
• Developing KPI Calculation reports on HIL and SIL playlist data with EMR Spark interacted with MongoDB on AWS
• Designed Architecture of ADAS application on AWS stack and Cloudera Hadoop - At the end of the project, I worked on database and application refactor migration from Cloudera to AWS cloud
• Vehicle Data Management API (PoC) - Design Vehicle Data schema on AWS DynamoDB
• Design and develop AWS Lambda REST API for GET, PUT and UPDATE data stored in DynamoDB
• Design Cloud formation templates for provisioning application platforms.

• Oncor Electric Delivery, North America Project (Live)
• Project Description and responsibilities: Oncor Electric Delivery Company is Texas's largest transmission and distribution electric utility.
• I worked on incremental loading of event logs, error log, and electricity meter data on Azure data lake
• Performed predictive breakpoint analysis on Smart meter analysis with Python and Spark on Databricks and Azure ML Service.
• Performed image classifications with computer vision on faulty network towers device predictive maintenance.

CTL-Cloudera Big Data As Service Project, USA (Live)
Project Description:
Project link: https://www.cloudera.com/about/news-and-blogs/press-releases/2016-09-21-cloudera- centurylink-expand-strategic-alliance-deliver-big-data-service-for-customers.html
BDaaS Project includes complex architecture including product knowledge, Hadoop services knowledge, Cloudera API knowledge, and Cloud automation which initialize CenturyLink
managed a secured cluster for Data Warehousing, Big Data Analytics applications.
Played an integral role as Subject Matter Expert of CenturyLink Big Data as a
Service cloud product. Responsibilities are to develop an Automation API framework in Java/Python which will set up and manage clusters with all services up and running
automatically.

Panera Bread Data Science Platform , USA (Live)
Project Description: Panera, LLC is an American chain of bakery-cafe fast-casual restaurants in the United States and Canada. CenturyLink has SOW with Panera, LLC for Capacity Planning and Production setup. The client required Identification of methodology for tying the online business workload at an order level to the actual utilization of the IT infrastructure and building of sample/representative dashboard/s depicting measures defining the IT resource utilization per order.

CTL Data Lake Project: Data Ingestion API (PoC)
Project Description: CTL Data Lake is a CenturyLink internal project for creating an
application for comprehensive data access and management and then applies data analytics on scalable data.
My responsibilities were to develop and test the REST interface for a data pipeline which
takes data from the customer and dumps to the Kafka topic as well as parsing it with Streaming and stored to the HBase table and HDFS.

DLL Financial solution partner, Mumbai, India (Live)
Design and Develop Azure Data Pipeline with Azure Data Factory, HDInsight, Azure Databricks, Data Lake Analytics, Azure SQL (Stage), Azure DW, Azure ARM and Power BI

Michaels Store, USA (Live)
Design Distributed high-performance Data analysis and Business Intelligence platorm on top of Cloudera Impala, HAproxy, etc.

A&T Insight is a module of CMS (Amdocs Customer Management Suite) is a low-latency Big Data and predictive analytics Applications on top of customer billing data to provide insight-based reports to customer care executives during a call with a customer. Responsibilities/Deliverables
• Software Development with Hive and PIG scripts for analysis of Huge Telecom data.
• Develop dynamic map-reduce jobs and UDFs with Core-java.
• Automatic Data ingestion platform for migration of data from Oracle and H-Base.
• Worked in distributed Gigaspaces Grid clusters for Insight Application.
• Exposure to real-time Spark stream jobs and batch jobs.
• Developed modules for database connections and for structured programming.
• Experience with both Hortanworks and Cloudera distribution of Hadoop (CDH).
• Exposure to data manipulation with Hive queries and Pig scripts on HBase data (No SQL)

• Developed log analysis and real-time monitoring tools for the production application.
• Exposure to ETL job creation, flow diagrams, and job scheduling in DAG.

AT&T Telegence Mobility project create and manage a set of ETL workflows with Pre-Data validation and Post-Data Validations jobs to generate customers' bills from Billing data.

Automation of Campaign mailing system
The campaign mailing interface was designed for the client relation team to test emails and trigger jobs to send a huge number of targeted emails.
- Worked on distributed 7-node cluster queues, Solr, and Sphinx framework.
- Have a good experience with Lucene queries with Mysql on a distributed cluster.
- Experience in MySQL Server 5.2 including the views, Indexes, Procedures Optimization and performance.
- Software development with Java, Python, Perl, Javascript, and MemCache in Live Projects.
- Experience in Advanced Automation programming in Python and Perl for Bulk mailing systems and system tests.
- Impact: Manual processing of the mailer campaign had been stopped with fully automated data product where data can be selected with just a saved query id of target customers and 200k to 2 million emails execution at a time with mail tracking features for demographic analysis of email readers. Awards: Team got awarded by APAC leaders.

Monster online job posts aggregator
Description: A Framework built for the reusability of the jobs scraping process to run 200+ processes at the time
Impact: 121,000 Job posts with this automation in a single month. which saves the workforce hours of monster.com clients.

Admin Mannual of GulfTrip website: A Real Time Web application is B2B process specifically developed to the process of providing accessibility,Sales Reports, Profitability Reports, and Business Rules forGulftrip Website.In this Application, the Client will be able to modify GulfTrip Website which is basically a Tourism Module Website developed with Company Loop Methods.
Work is Development, Testing, and Support
Impact: long contract from client

• The role includes Development, Testing, and Support of the website portal of the Admin Manual of GulfTrip website: A Real-Time Web application is a B2B process specifically developed to the process of providing accessibility, Sales Reports, Profitability Reports, and Business Rules for Gulf trip Website
• In this Application, the Client will be able to modify GulfTrip Website which is basically a Tourism Module Website developed with Company Loop Methods.