SUMMARY

• Extensive experience in RPA Development with the focus in Banking, Financial, Retail, Transportation, and Telecom
• End to end RPA life cycle design and development expertise using WorkFusion Machine Learning - SDK, OCR, SPA, RPA Express, and Pega Robotics
• Successfully deployed several Cognitive bots which include connecting with different servers such as AS/400, APIs and use of different Java libraries such as, EWS, iText
• Created ML workflows which involve training large data sets for WorkFusion Machine Learning Models
• Developed OCR (Optical Character Recognition), Information Extraction, Classification & Data Categorization for several projects
• Designed solutions that connect a Chatbot to WorkFusion RPA technology through API connections
• Configured and deployed Excel Macros to automate heavy data entry related use cases
• Implemented processes that monitor emails, download pdf documents, perform OCR, and tag data by Machine Learning models and upload them to AS/400 server, this reduced manual document handling errors and handling time by 75%
• Executed Cognitive Bot with OCR, Machine Learning and RPA which deals with the substantial amount of unstructured data handling, saving processing time by 80%

TECHNICAL SKILLS

RPA Tools: Work Fusion, Pega Robotics

Work Fusion: ML, Control Tower, Scheduler, SPA Studio, RPA Express, Workspace, Mesos, Marathon, Nexus, AWS, Amazon S3, ML-SDK

Programming: Java, Groovy, XPATH, JavaScript, Maven, Selenium, .Net, C#, HTML, XML, JSON, CSS, Visual Basic, Hibernate, jQuery, J2EE, SQL

Other: Git, Web Service, Eclipse, Visual Studio, API, Tableau

PROFESSIONAL EXPERIENCE

WF ML Engineer & RPA Lead Consultant

Confidential

Responsibilities:

• The Accounts Payable team of this largest retailer in Canada handles over thousands of supplier’s invoices which comes through emails daily
• SME's input data from the invoices to PeopleSoft application manually
• An RPA bot downloads the vendor invoices via Exchange Web Service (EWS)
• A machine learning model processes those invoices (in PDF format) by extracting details from those
• A second RPA bot logs into PeopleSoft and inputs data in the form to create vouchers and reports
• OCR failed to process correctly for some image-based invoices which caused the model to extract wrong data
• Created post-processing rules to capture those invalid data and route those invoices to the workspace to be tagged manually
• The Finance department moves emails manually to different email folders to be looked at by the associated teams
• These emails are unstructured and don’t contain a consistent identifier to implement Outlook rules on them
• A machine learning algorithm is used to train the model which reads the body of an email, identifies the pattern and classifies them into a predefined category
• Based on that category rule-based bot moves them to their corresponding folders
• Due to the varying nature of the email contents, it was difficult for the model to classify them correctly
• Retraining the model with new types of emails and large dataset reduced the errors

Environment: WorkFusion, Java, ML-SDK, PeopleSoft, Exchange Web Service (EWS), OCR

Confidential, ON

Responsibilities:

• A leading LTL transportation company in Canada which does cross-border shipment of 1000+ orders every day through an email
• To process these shipments through the border, the client sends documentation as a customs broker package with certain information to the customs which goes through a customs broker for further processing
• Invoices come in multiple formats - pdf, excel, tiff or png and bundled with several other documents such as NAFTA are not required for the final package
• Used iText to convert all file formats to pdf for better OCR, splits pdf so that other documents can be discarded
• Added rule-based classification to identify between invoices and other documents, merge the invoices only for the information extraction step
• Invoices contain data which scattered all over the place. Orders come with different templates that made it difficult for the model to extract data from one place
• Trained SMEs to tag gold data from a consistent place for better learning of the bot, trained bot with large datasets around 800+ for more accuracy, connected to the client’s API to scrape more information in case bot missed any at the extraction step
• One of the leading financial institutions in Canada required mortgage insurance validation for loans with a down payment of less than 20%.
• This institution processes a large volume of documents manually every day. Documents come via email or post and then got uploaded in the AS/400 server manually, then matches the customer information to identify the correct customer to update the insurance status
• Developed a solution where the bot monitors the email server, downloads insurance letters, uploads them to the AS/400 server, also send the document to the OCR queue where the documents got processes for information extraction, customer's data got extracted by the Machine Learning bot by automatic tagging for further processing. RPA was programmed to connect to the CRM system to find the correct customers by providing the data from the extraction step and update the database with the insurance status
• If insurance was cancelled the bot sends an email to the customer to take necessary steps
• Improper scanning of insurance documents, OCR was providing poor quality documents which made it difficult for the Machine learning bot to tag data from them. Converted pdf to tiff, added correct OCR skew to handle poor quality documents which provided better results
• Uploading files to AS/400 were failing due to the time lapse between bot and response from the server
• Added delay for the bot so that it can connect efficiently with the server
• Provided WorkFusion rule-based solution to a leading telecom company in the USA for the procurement process.
• This use case involved a large amount of data entry between Excel files and CRM systems
• If any information was missing from the dataset, bot used to get stopped until it gets manually updated
• Added logic to handle these exceptions so that bot can process the next line items and sent an email to the SME with the exception order number for the additional steps

Environment: WorkFusion, Java, iText, AS/400, ML-SDK, CRM

Confidential, ON

Responsibilities:

• Implemented a solution for post-implementation exception handling for the production support team
• Was responsible for server maintenance, bot scheduler, and infrastructure management
• Provided SME’s with PEGA Robotics development training to handle the post-production issues

Environment: Pega Robotics, C#, .NET

Confidential, ON

Responsibilities:

• Designed and developed RPA and RDA bots using Open Span/Pega Robotics and C# in the .NET platform for the bank’s contact centre
• Developed the connectors between the bot with several legacy applications via interrogating the elements in the web and desktop applications, scraping data from HTML tables and presenting them in a User Interface for the agents to process the customer’s request quickly
• Created Accessible User Interface, interaction manager components to connects all customer interaction through the configuration of its properties, events, and methods, so that existing applications are unchanged, and the bot can fetch information through the interrogated fields
• Created workflows as per the process maps, technical design documents and test cases for UAT to ensure a seamless delivery of the quality product
• Introduced Kanban methodology to ensure the continual delivery of the processes and better integration to provide a quality product

Environment: Pega Robotics, C#, .NET

Confidential, ON

Responsibilities:

• Monitored the performance of web clients using LoadRunner and SOAPSonar to ensure RBC’s web applications’ smooth functioning
• Documented processes of accessing RBC’s most secure retail application “Sales Platform” to assist non-technical team members with understanding the techniques quickly and working efficiently
• Created automation test scripts using C# and .NET based on business requirements to ensure the quality delivery of software

Environment: LoadRunner, SOAPSonar, C#, .NET

Confidential, ON

Responsibilities:

• Charged with testing the accessibility of web and desktop applications using JAWS screen reader and Color Analyzer against 20 accessibility guidelines in WCAG 2.0; documented issues and advised developers on possible solutions
• Developed video caption authoring tool using C# and .NET for hearing-impaired viewers in RBC which helps them learn the training materials effectively
• Re-designed IT Accessibility’s internal social media page for retaining current users by incorporating a better content strategy

Environment: JAWS, WCAG2.0, C#, .NET

Confidential, ON

Responsibilities:

• Assisted 40,000+ students, alumni, and staff members in resolving technical issues including software, hardware, network, and email issues and threats; utilized remote assistance tools such as the VNC server
• Logged 300 accurate and descriptive support tickets daily through the PeopleSoft CRM systems

Confidential, ON

Responsibilities:

• Worked alongside a team to develop a digital solution that saved the company $1,000 per month in inventory-related expenses; discovered the root cause of the inventory complications, drafted use cases and activity diagrams, and devised a solution to the problem
• Ensured compliance of inventory processes and procedures on 20,000+ merchandise items by regular firm auditing

Java Developer

Responsibilities:

• Developed an inventory management system using J2EE and SQL to ensure real-time product availability for the online store
• Was responsible for implementing updates and patches to tune the system to achieve optimum performance level and provide smooth system integration
• Assisted with detailed technical design process included initial testing, development, implementation and maintenance

Environment: Java, J2EE, SQL