Jun 2024 – Dec 2024 · 6 mos · NJ · Remote

• Worked on GenAI project.

Sep 2020 – May 2024 · 3 yrs 8 mos · New Jersey, United States

• Client: KPMG, EY, Nowigence and Other clients
• Project: Data Workbench and Workflow Mining, Industry & Technology Risk Intelligence, AI Proctoring (EAssesMe), Pluaris
• Tools: Python, TensorFlow, Kera, PyTorch, Spacy, Transformers (Roberta, Bert, Bigbird and Distilbert),
• semantic search, MongoDB, Plotly, and Flask, Yolo, AI Service (Face Recognition), Voice Algorithm and Data Annotation,, LLM Stack, LLM model, and Azure Cognitive Search, ADF, SQL, ADLS,
• Databricks, Azure Synapse, Power BI, Azure ADO.
• Analyzed the existing products and market trends to propose product roadmaps to the senior
• executives and customers to set short- and long-term goals. Created the project execution plan and
• managed a team of data scientists/statisticians and software engineers for the development of pluaris
• platform. Responsible for successful project closure.
• Responsible for data strategies like data collection, quality, compliance, governance, and management.
• Worked across Computer Vision / NLP / ML/ IR Models to a broad-based / End-to-End Analytics with AI
• driven cognitive decisioning platforms/products, business critical domain-specific solutions, and Big Data
• engineering services.
• Architected, developed, and deployed the NER, Text Summarization, Chatbot, Semantic search, Fraud
• detection, Demand & Sales Forecasting , Near-Realtime IoT Greenhouse monitoring & Recommendation
• system.

Feb 2020 – Sep 2020 · 7 mos · Greater Chicago Area

• Clients: Mobilize and AstraZeneca
• This project focus on developing a prediction model for the mining company to identify the defects and drift in the drilling operation (IoT/Sensor Data). We use customer call text data and provide the sentiment on customer service representatives.
• Responsible for Data dictionary, data mapping, SQL development for data quality reports.
• Developed ML/DL (sage maker) models and automated the deployed with CI/CD pipeline in the AWS cloud and other data services. ML/DL: Computer Vision (TensorFlow) and Sentiment Analysis (NLTK, Spark MLlib).
• Wrote and tested Python scripts to create new data files for Linux server configuration using a Python template tool.
• Created a Handler function in Python using AWS Lambda that can invoke when the service is executed
• Developed the MapReduce jobs in Python for data cleansing and data processing
• Team to build a Data Mata for building guardrails; used SQL scripts to bring the data into the Data Mart
• Working on the Salesforce Einstein Analytics platform.
• Used the Sqoop to import the data from MySQL and SQL Server into the Hadoop Ecosystem.
• Involved in the analysis and design of change requests through the logistics tracking system. Built SQL stored procedures and triggers.

May 2019 – Feb 2020 · 9 mos · Bengaluru, Karnataka, India

• DDL project focus on developing an end-to-end data pipeline (bigdata architecture and DL prediction models) for Lam research (semiconductor company) enterprise application to identify the defects in their production process (IoT Devices/machines).
• Created the ETL/ELT pipelines to import the data from various sources, processed and loaded them into Data Lake.
• Collected data from variety of sources including NoSQL and relational databases, various files and web services as needed using Flume and Kafka
• Job scheduling, batch-job scheduling, process control, forking and cloning of jobs and checking the status of the jobs using shell scripting.
• Managed datasets using Panda data frames and MySQL, queried MySQL database queries from Python using Python -MySQL connector MySQL db. package to retrieve information.
• Developed and deployed Multiple Scalable Deep Learning Models which can produce results in less than 3 secs (SLA) in the Enterprise product.
• Developed AI & Deep learning models (Computer Vision, NLP) as microservices (API).
• Designed the entire Bigdata processing pipeline includes (Redshift, Hortonworks- HDFS, Spark, Kafka, HBase, Phoenix), SQL, and ELK (Elastic Search).
• CI/CD Deployment - in Dev, Test, stage, and prod environment (Containers, Jenkins, Git).
• Research and benchmarking on the best practices for our product development and deployment.
• Practices scrum meeting.

Dec 2017 – Apr 2019 · 1 yr 4 mos · Chennai, Tamil Nadu, India

• Develop the Health Information Exchange platform, which aggregates healthcare data (HL7, text, and other standards) from multiple sources and provides insight using the platform.
• Managed 5-member team and practice scrum
• Collected external text data from web sources and perform text processing and analysis to identify the population health or disease outbreak.
• Architected the production system (data platform) which can cater to 1 million customers.
• Developed healthcare analytics data ingestion platform.
• Working on Bigdata technologies (Hadoop ecosystem, - STORM, KAFKA, HIVE, PIG, FLUME, Presto, SPARK, and Oozie)
• Wrote Python routines to log into the websites and fetch data for selected options and used Python modules such as requests, urllib, and urllib2 for web crawling.
• Prepared and analyzed reports using Python libraries and involved in environment Setup.
• Created the environment-specific settings for a new deployment and update the deployment-specific conditional checks in the codebase to use specific checks.
• Reviewed basic SQL queries and edited inner, left, and right joins in tableau desktop by connecting live/ dynamic and static data sets.
• Involved in reviewing business requirements and analyzing data sources from excel/ SQL server for design, development, testing, and production rollover of reporting and analysis projects within tableau desktop.
• CI/CD Deployment, Dev, Test, stage, and prod environment (Bamboo, Docker, Maven/sbt, Git/Bitbucket)
• Designing the entire Bigdata processing pipeline in the hybrid cloud environment (AWS and VMware) for conducting experiments. Architecture - Nginx, EC2, API Gateway, S3, and others.
• Research and benchmarking on the best practices for our product development and deployment.

Sep 2017 – Dec 2017 · 3 mos · IBM T.J WATSON RESEARCH, NY, USA

• Multi-agent-based Swarm Intelligence AI
• Developed a multi-agent-based swarm intelligence algorithm for decentralized power token and information exchange problems.
• Accomplishment:
• Identifying and developed a Multi-agent based Swarm Intelligence AI simulator for the research problem.
• Developed tools using Python, Shell scripting, XML to automate some of the menial tasks.
• Identifying and proposing algorithms to the research problem (Python) and IBM Watson Platform.
• Evaluating the proposed architecture, algorithm with the existing system, and Publishing the results.
• Benchmark performance and optimize the Power token exchange and communication delay in Power 8.
• Reporting Manager (mentor): Dr. Pradip bose (Manager, Resilient Energy Group), IBM T.J Watson Research, Yorktown, NY, USA

Jul 2014 – Dec 2017 · 3 yrs 5 mos · Anna University, Chennai

• Anna university is one of the top research universities in India. My research proposal addresses the problems related to the smartcity project. We are developing a platform for addressing the environmental and traffic related issues as a part of smart city initiative in order to provide the eco-friendly environment. It’s a multidisciplinary research which includes bigdata, cloud computing and Internet of Things (IoT). We use the sensors as data feeds (or multiple sources) to analyze and perform both near-realtime (edge level) and batch analytics (cloud level) that can provide valuable insight to mitigate the above problems.
• Responsibilities:
• Designing the entire Bigdata processing pipeline in the hybrid cloud environment (Microsoft Azure
• and IBM Bluemix and private cloud) for conducting experiments.
• Working (contribution) on improving the bigdata frameworks like hadoop (data partition skew) and spark (GPU and effective utilization system resources- mainly in POWER systems).
• Working on Bigdata technologies(Hadoop, STORM, HIVE, PIG, FLUME, KAFKA, SPARK) and NoSQL databases(MongoDB).
• Analyzing the state-of-the-art algorithms and approaches, and identify the gaps.
• Working on developing novel algorithms and create models using R (RStudio) and Machine learning.
• Deployed on-premise private cloud in blade servers using openstack.
• Projecting the results using Data Visualization techniques (Tableau).
• Developing IoT prototypes using Raspberry pi, Arduino, and environmental sensors.
• Achievements:
• We got funding from IBM smartplanet initiative of worth $5000 and IBM Bluemix platform for our project.
• Our proposal got selected as finalist in Anveshan IoT Student fellowship Project, funded by Analog Devices.
• Funded by Microsoft Azure Research (2015 - 2016) by providing their platform to conduct experiment.
• DST-PURSE funded consumable grants of worth Rs 2 Lakhs.
• Team Size: 9 Member (Team members : 1 PhD , 3 Masters, 5 Bachelor students)

Jan 2011 – May 2012 · 1 yr 4 mos · Trento Area, Italy

• Research Area: Peer-Peer and Cloud computing
• Domain: Distributed Systems
• Institution Involved ( Research Universities),
• University of Trento, Trento, Italy
• Indian Institute of Madras, Chennai(Madras), India
• Guide,
• Prof. D. Janakiram , IIT Madras
• Prof. Alberto Montresor , University of Trento.

Feb 2010 – Jun 2014 · 4 yrs 4 mos · Chennai Area, India

• ICSR is an industrial consultancy located in IITM. I worked in different research projects from Thales, Xerox and DST, and lead the team with different size based on the requirement. My role is to design and develop software, and guide the team (4 to 6 members). Finally, I am responsible to report the progress of my team to the head consultant and client’s along with deliverables. I worked in bigdata and cloud computing projects.
• As part of our OI bigdata project, we aggregate different types of data from multiple sources to create a common framework to answer the business questions or problems. We addressed problems due to huge volume, noisy nature of the data and scalability issues. We propose a framework which provides Operational Intelligence (OI) - categorization and prediction to provide insight into business operations by running query analysis against offline and online data with minimal overshadowing effect (Tradeoff based). It reduces the business negative impact during the customers product launch.
• The goal of Clive is to provide P2P Live broadcasting by using cloud computing infrastructure to guarantee a minimum level of QoS. In this project, we saved up to 45% of the cost by choosing the right number of active helpers (cloud front and EC2) as compared to the previous version which uses only passive helper (S3) to guarantee the predefined QoS.

Jan 2008 – Apr 2009 · 1 yr 3 mos · Anyang, Gyeonggi-do, Korea

• Work Nature:
• Worked as Offshore coordinator.
• Involved in Middleware Development
• About Project:
• Goal of this project is to develop a Universal Device Middleware (Device Driver) for our Industrial Automation product so we developed a Common architecture for handling all Device with different functionality which uses different Protocol stacks like TCP/IP, UDP/IP, All serial standard Protocols and ModBus TCP/Serial/UDP.
• This Middleware act has a controlling Layer for all our device application by using FDT/DTM Technology. The major role of Middleware is to provide the following functionalities like, Protocol Finder, Device Manager, Debugger, Functional Component Layer, Data mapping and RAD Tool.
• Responsibilities
• Complete Project management in onsite. ( worked along with Project Manager)
• Experienced in complete Product Development.
• Optimizing and classifying the device drivers
• Coordinating with Offshore development team
• Reporting Project manager in Regular Intervals.
• Tools used
• Project management : Microsoft Project
• Design Tool : Visual Studio Class Designer, VISIO
• Software Used : C# 2008, C/C++
• Work Domain : Embedded, Industrial Automation (DCS, HMI, Device Driver )
• Protocol : RS232/485, TCP/IP, UDP, MODBUS.

Jun 2007 – Dec 2007 · 6 mos · bangalore

• Tata Elxsi is one of the major service providers in the consumer electronics area.
• Developing STB (Set-Top Box) using OpenTV middleware in the EMMA architecture board.
• Developed functional components using OpenTV, my role is to develop HAL (Hardware Abstraction layer) between Device driver and Middleware which act has a Mapping layer by implementing the middleware interface.
• Identifying the suitable middleware interface for implementing Hardware Abstraction Layer from the OpenTV specification.
• Implementation and mapping the functionalities between two different layers.
• Testing and assisted in deploying on EMMA architecture board (STB board)

Apr 2006 – Jun 2007 · 1 yr 2 mos · Chennai Area, India

• Project: UMU Antivirus:
• Goal of this project is to develop a Mobile Anti Virus Product for UMU Ltd. This UMU Scanner is an Antivirus product for wireless ARM devices like PDA's and Smart phones. UMU Scanner has three modules Scan Engine is used to find the appropriate virus and virus infected file in the wireless device.
• Quarantine will make the virus and virus infected files into inactive state then it will restore the infected file after removing the viral patterns from the file. In update, we can download file from the update server. There are many modes to download updates from the server; it can be achieved through GPRS,
• Wi-Fi and through Active Sync.
• Mobile Tracker was developed for a windows mobile app for tracking your lost mobile.
• Responsibilities
• . Worked as Module Leader (Managed 3 member team)
• · Active Participation in Product Development.
• · Keep track project execution, manage changes, develop and execute implementation plans in stipulated time and deploying the solution.
• · Reporting to the Team Leader and CTO in regular interval.
• · Coordinating the team members.
• · Coordinate with other internal departments including Graphics and Testing.
• Tools used
• Project management : Microsoft Project
• Domain : Mobile & Embedded
• Software Used : C, EC++, C#.Net

Jun 2002 – Jun 2003 · 1 yr · Chennai Area, India

• We worked on a project which deals with transmission of media file from server to client without any time delay using RTP, RTSP, and RTCP. We mainly focus on Audio (mp3) file streaming from the server, which is both played and streamed to all connected clients at the same time. Media file will be received, played and stored at client machine without any time delay.
• The major advantages of multimedia streaming system are no delay and jitter free. Here RTCP Protocol used to control the Stream of information from the server to client.
• Responsibilities:
• · Go through the design documents and discuss with Team leader regarding the design and client requirement.
• · Implementation document and implementation of project
• · Document my work and report to team leader.
• · Unit testing and preparing the bug report.