Sandip Jha

Experience

• 

  Infinite Computer Solutions

  Jan 2005 - Apr 2006

  Team Member

  Was involved in testing for Corporate and Number Portability project

• 

  Qwest Telecom

  Jan 2006 - Jan 2007

  Software Engineer

  Was involved in test case preparation and execution.Defect Reporting and retest.Attending client call.

• 

  Tata Consultancy Services

  Dec 2007 - Mar 2017

  At present involved in Functional Testing for Client ME Bank(AUS), where legacy systems are being replaced by T24 (R13). Involved in Lending and Collateral module managing SIT Functional Testing from Pega to T24.Attending BRD and FSD walk-through,analysis of BRD and FSD,developing test design, Execution of Test cases, coordinating defect triage,assisting UAT team,preparation of status report,Knowledge sharing session. I was involved in testing for modules like Customer,Payments and Asset Management for Client DB PWM IT where T24 was upgraded to R11. Worked for client JPMC UK a treasury project implemented on T24. I was involved in following modules Customer,Account,Interest ,Statements and various tax reports. It was a CARDS project of JPMC, where I was leading CSD division a testing team of 10 associates. We were responsible for testing Payment system. Short and long term assist payment program to offer defaulter customer.

  • Test Lead

    Aug 2012 - Mar 2017

  • Team Manager

    Nov 2011 - Aug 2012

  • Team Member

    Aug 2010 - Oct 2011

  • Project Lead

    Dec 2007 - Jul 2010
• 

  Capgemini

  Jun 2017 - now

  Currently managing a GenAI LLM testing project for clients in the hospitality industry.Client's Smart Chatbot on GenAI LLM Model (gpt-35-turbo-0125 and gpt-4 with RAG)Your Role, Company Name - Dates of InvolvementProject Overview: Engaged in a pivotal project utilizing GenAI LLM models to develop Customer's Smart Chatbot, aimed at enhancing business operations through advanced AI-driven data handling and response capabilities.Scope:Ensured successful document uploads and implemented semantic and vector search capabilities.Validated on-demand vectorization of documents triggered by user prompts.Tested the AI Assistant's accuracy in responding to diverse business-related queries within Org.Assessed the chatbot’s adherence to non-disclosure of its operational prompts and rules under multi-prompt scenarios.Confirmed the integration and accuracy of Optical Character Recognition (OCR) for varied document types.Verified the chatbot's ability to handle multiple queries within a single input and recognize key entities accurately.Automation Approach:LLM Model API: Developed an automation framework using OpenAI’s API to test response accuracy and relevance, integrating semantic similarity assessments to ensure conversational precision.API Testing with RAG Model: Implemented a Python script to automate interaction with Microsoft Bot Framework's Direct Line API, simulating user interactions and capturing chatbot responses for performance evaluation.Enhanced the chatbot’s response capabilities by integrating Azure Cognitive Search, improving information retrieval and document reference accuracy, showcased by dynamic reporting in Excel with performance metrics and outcome visualizations.Based on Prompt Engineering & Parameters tuning different scenarios can be explored 1.Profile & General Capabilities 2.Output Format 3.few shots 4.Instructions Adherence 5.Gathering and Presenting Information 6.Contextual Understanding 7.Response Temperature and Top P 8. Index Show less

  • Senior Manager

    Dec 2023 - now

  • Manager

    Jun 2017 - Nov 2023