Dr. Pandurang Kamat is Chief Technology Officer at Persistent Systems, he is responsible for advanced technology research focused on unlocking business value through innovation at scale. He is a seasoned technology leader who helps customers improve user experience, optimize business processes, and create new digital products. His vision for Persistent is to be an innovation powerhouse that anchors a global and diverse innovation ecosystem, comprising of academia and start-ups.
Pandurang joined Persistent in 2012. Prior to Persistent, he was the Director of Analytics for Ask.com’s search and content businesses, where he led a global team to manage Ask’s analytics platform. Before that he helped build secure communications and digital media products at Bell Labs and HP Labs and an award winning wireless research platform at Rutgers University.
Persistent Systems is a trusted Digital Engineering and Enterprise Modernization partner for global market leaders across Industries.
What initially attracted you to computer science and computer engineering?
My interest in computer science and engineering was sparked during a summer course in school. Learning programming constructs and creating computer games introduced me to the structured logic that supports these fields. I was captivated by the ability to break down complex problems and solve them systematically. What truly drew me in was the immense leverage that well-designed programs offer. They can automate tasks, optimize processes, and empower individuals or small teams to achieve remarkable feats. This blend of creativity, problem-solving, and transformative potential continues to inspire me. From those initial experiences to my ongoing journey, I remain passionate about the endless possibilities that technology presents. Computer science and engineering not only shape the future but also offer avenues for innovation and progress that drive me forward.
The bulk of Persistent Systems business comes from building software for enterprises, how has the advent of generative AI transformed how your team operates?
The advent of generative AI (GenAI) has transformed how our team operates at Persistent, particularly in enterprise software development. This disruption within the IT industry not only presents challenges but also significant opportunities to reimagine business operations holistically.
As an AI-powered Digital Engineering enterprise, Persistent has embraced GenAI to revolutionize various aspects of the software engineering lifecycle. Over the past year, we have developed tools and suites that completely redefine processes such as code generation, test case generation, and report migration. In legacy modernization projects, our approach has evolved significantly. We now leverage tools to streamline code takeover processes, mitigate project risks, and expedite the onboarding of new team members by providing them with a deeper understanding of complex codebases. Furthermore, our collaboration with industry domains enables us to deliver tailored solutions leveraging enterprise data. By developing digital assistants capable of understanding business language and providing relevant references, we enhance operational efficiency and decision-making within enterprises. These assistants adhere to Responsible AI principles, ensuring transparency, accountability, security, and privacy while continuously improving their accuracy and performance through automated evaluation of model output.
What are some of the challenges of completely modernizing legacy systems using generative AI?
GenAI is a powerful tool, but it’s not a silver bullet for complete legacy system modernization. Organizations across industries must adopt a combined approach, harnessing human expertise and AI capabilities. While GenAI offers substantial potential for modernization, it has its limitations. Key challenges include:
- Limited Understanding of Legacy Systems: GenAI models require a thorough understanding of existing systems to function effectively. Legacy systems often lack comprehensive documentation, hindering the ability of AI to grasp their interdependencies effectively.
- Data Quality and Bias: The quality and representativeness of data used to train the AI model have a significant impact on its output. Limitations of the training data can be reflected in the generated code, potentially introducing new problems.
- Ensuring Quality and Security: While GenAI can automate code generation, the output needs rigorous testing and verification to meet quality, functional requirements, and security standards.
- Limited Scope of Modernization: GenAI may be unsuitable for complete system overhauls. It can excel at specific tasks like code refactoring or test-case generation, but complex architectural changes still require manual intervention.
- Change Management and Stakeholder Alignment: Managing organizational change and gaining stakeholder buy-in are critical factors in determining the success of modernizing legacy systems with GenAI. Clear communication, training programs, and stakeholder engagement initiatives can help address resistance to change and facilitate smooth transitions.
One of the challenges of Generative AI is consistency, how does Persistent Systems assist with building a consistent user experience?
Consistency is one element of providing an overall enterprise-grade, enterprise-safe GenAI-powered user experience and outcomes. We look at the process holistically.
We provide end-to-end support across all stages of GenAI adoption. Our strategic guidance and meticulous use case analyses aid organizations in selecting the most suitable foundation models (FMs) tailored to their specific requirements. Through a detailed examination and consultatn, we assist clients in defining clear use cases and making informed FM selections.
Then, we focus on multiple approaches, such as few-shot prompting or even fine-tuning, to ensure that the models used in the applications are attuned to the use case and enterprise data.
Our solutions not only employ standard RAG techniques but also go deeper into multiple prompting and data chunking strategies to ensure the most relevant data is retrieved and given to the FM during inference. We further enhance the accuracy and relevance of this context by using advanced Knowledge Graphs to capture hidden relationships within the enterprise data.
We also employ multiple grounding techniques and guardrails to limit and focus the purview of inference.
Finally, we put the application through a rigorous and automated evaluation framework that ensures consistency of inference and experience, release after release.
Could you provide real-world examples where GenAI-powered solutions have successfully revolutionized customer interactions?
Persistent has transformed customer interactions for a leading software solutions provider through GenAI-powered solutions. Facing scalability challenges during peak operational periods, the company implemented a Central Knowledge Repository and Conversational AI Teams BOT. It streamlined access to information, leading to 80% reduction in customer query resolution time. The quality of responses also improved significantly, resulting in enhanced customer satisfaction.
We also assisted a private equity firm by leveraging GenAI to automate the creation of detailed investment reports. With the GenAI-powered system, the time required to generate reports was reduced by 90%. This streamlined approach revolutionized the firm’s operations, facilitating rapid and effective decision-making. The efficiency not only saved valuable time but also fostered increased collaboration among stakeholders and ensured a personalized touch in each memo, enhancing overall effectiveness.
How do you approach Responsible GenAI innovation?
Our approach to Responsible GenAI innovation prioritizes ethical practices and regulatory compliance throughout the development and implementation processes. We emphasize transparency, accountability, and fairness in AI-driven decision-making.
We establish robust ethical guidelines governing the development, deployment, and use of GenAI systems. In our pursuit of Responsible GenAI innovation, we rigorously test and validate our systems to mitigate potential risks such as biases, misinformation, and privacy issues.
Furthermore, we prioritize transparency and accountability in AI-driven decision-making processes by providing users with clear insights into system operations. Ultimately, our approach aims to develop and deploy GenAI systems that drive innovation and efficiency while positively contributing to society.
What is your vision for the future of AI?
My vision for the future of AI is multifaceted. Firstly, in digital engineering, I envision AI not only as a coding assistant but also as a collaborative partner, similar to a “pair programmer.” This involves AI assisting in coding tasks and actively participating in problem-solving by mapping out complex tasks and executing sub-tasks.
Secondly, I foresee an era of personalized AI agents and assistants offering tailored experiences to individuals – a “personalization of 1” approach. These agents will understand users’ unique preferences, behaviors, and needs, providing highly customized support and services.
Lastly, I believe in the evolution of compound AI systems, where various AI models coexist to address different needs. There won’t be a single “one-size-fits-all” model, but rather a combination of large and small, general, and purpose-built models working together in AI services. This approach allows for greater flexibility, efficiency, and effectiveness in solving a wide range of problems across different domains.
Thank you for the great interview, readers who wish to learn more should visit Persistent Systems.
Credit: Source link
