Kamal Rajendran

• Creating an ML web application to train and retrain a model on different data sets to find the defects on semi conductor wafer chip.
• Involved in Design for the product.
• Developed end to end product.
• Achieved 95% of accuracy.
• Successful product in Samsung and Micron

Packages used:

• Cx-Oracle to connect oracle db and fetch data.
• Bokeh, Holfffoviews and Datashader for Data Visualization
• Sklearn for creating models
• FastApi for receiving data with schema definitions to view metrics chart to check the performance of the model
• Building Docker images for application scalability
• Configuration of pods for containerization on single node cluster as services and configuring services in ingress to access on specific api
• Maintaining the code in azure git repository
• Continuous integration and continuous deployment using azure CI/CD pipeline

Client: KLA

• Worked as data engineer where I had involved in migration of historical and incremental data from different data sources from AWS.

Services:

• (DynamoDB, RDS and S3 Bucket) into snowFlake database.
• AWS Glue Workflow which runs as batch job to transform data in order to maintain source and target system as up to date

Actively involved in creating cloud formation templates for the deployment.

Configuring S3 notification with SQS events and Snow pipe,

ensuring smooth data flow from s3 to snowflake.

Developed a python script within Lambda for efficiently fetching stream data from Dynamo DB, subsequently writing into s3 bucket for seamless integration.

Involved in implementation, review, monitoring, documentation and maintenance of the application

• Client: Caterpillar Inc

• To develop a tool which create web application code with given properties such as language (NodeJS) and Database (Mongo/SQL)
• Write backend service code details such as language, database parent child properties, CRUD functionality services like create, update and delete in JSON format
• Our code will read JSON file and create application code in required manner
• Created a generic CRUD functionality code in EJS template form of NodeJS language, which generates whole backend service code with given properties
• After successful creation of web application code, we run unit test cases for all CRDU operations
• Implemented some customized technique in code creation template on database linkage between entities like foreign key relation
• Client: iCETS (Infosys center for Emerging Technology Solutions)

• To develop a tool which convert QuickJob to Cobol and Easytrieve to Cobol
• We will be giving input directories of QuickJob and Easytrieve code from user defined windows application and converted COBOL code, output will be stored in the respective directory given in the same application
• Downloading the required WxPython modules from Python library using PIP
• Respective QuickJob and Easytrieve keywords will be stored in a separate file
• Processing the input files by reading the data and replacing with the respective COBOL keywords
• And store them into the output directory
• Client: iCETS (Infosys center for Emerging Technology Solutions)

• Need to fetch all details about Medicine like name, usage, disease treated by this medicine, their ingredients, side effects, etc
• From unstructured file formats in a different way based on business requirement and generated a report in the form of report(Excel) and also store information in MongoDB from client machine
• Identify the file format from different folder and extract data from identified unstructured file formats (like pdf, docs, docx and image) with the help of Infosys internal tool (IECP - web application) and process in such a way to fulfill their business requirements
• IECP tool was developed in python language with Django framework and has lot of data extraction services as well as machine learning techniques
• Implemented logger exception in IECP tool, which helps to debug the error in minimum amount of time
• Generated customized models for MongoDB in IECP tool
• Implemented a technique to find PDF header and footer from the given documents in IECP tool
• In addition, we have implemented data extraction in many languages that supports in IECP tool
• Client: Novartis