/dq/media/agency_attachments/UPxQAOdkwhCk8EYzqyvs.png)
/dq/media/member_avatars/WhatsApp Image 2024-05-15 at 1.47.33 PM.jpeg )
Follow Us/dq/media/media_files/rg2iZbx3RV1UN2Z0i0Jr.jpg)
In a conversation held during the Confluent Kafka Summit, Kiran Kumar Katreddi, Vice President of Platform Engineering at Meesho, shares valuable insights into leveraging generative AI to revolutionize e-commerce experiences. Meesho, India's leading social commerce platform, has been at the forefront of incorporating cutting-edge technologies to democratize e-commerce for users across diverse markets. Here are the key takeaways from the interview:
How are you leveraging generative AI while considering its impact on your workforce?
Kiran Kumar Katreddi: We've been exploring various use cases for generative AI, particularly focusing on customer payments and enhancing user experience, all while keeping our workforce in mind. Our goal is to democratize e-commerce for users across different cities and empower our employees with these new tools. We're working on improving the end-user experience during purchases. For instance, we're experimenting with virtual try-ons for products like T-shirts and exploring other use cases to personalize user feeds based on their preferences using language models.
Can you elaborate on the challenges and opportunities you faced while integrating generative AI, specifically regarding your workforce?
Kiran Kumar Katreddi: Availability of GPUs for training models and managing the hype versus reality were significant challenges. We've been working on this even before it gained widespread attention. It's crucial to ensure a positive ROI and scalability. Cost efficiency is vital, especially in a price-focused market like ours. In terms of workforce impact, we view generative AI as an opportunity to upskill our employees and make them more productive. We're focusing on providing them with the necessary training and tools to collaborate effectively with AI.
How does Meesho ensure ethical AI practices and handle biases while considering the workforce?
Kiran Kumar Katreddi: We prioritize data governance and minimize biases by controlling data inputs. Our data science team ensures that only relevant and ethical data is used. We avoid external data sources to maintain objectivity and focus on enhancing user experience ethically. This approach also protects our workforce from potential biases that could be introduced through external data.
How do you measure ROI and success metrics for AI deployments, considering the workforce impact?
Kiran Kumar Katreddi: We tie AI initiatives to business metrics and conduct thorough experimentation before full deployment. Monitoring business metrics alongside control metrics is crucial. We scale up gradually, ensuring positive impacts on business goals while also measuring how AI is affecting employee productivity and satisfaction.
What advice would you give to companies incorporating generative AI, focusing on workforce considerations?
Kiran Kumar Katreddi: Start with small, domain-specific models aligned with business goals. Experiment, measure impact, and scale gradually. Focus on tangible outcomes rather than deploying AI for the sake of it. Ensure a clear understanding of the business metrics you aim to influence. When it comes to your workforce, prioritize upskilling and focus on collaboration between your employees and AI tools.
Could you name some tools Meesho uses for generative AI?
Kiran Kumar Katreddi: We use platforms from Google and Microsoft, along with our in-house ML platform, for for generative AI tasks.
How did the incorporation of generative AI impact your workforce?
Kiran Kumar Katreddi: We haven't significantly increased manpower. Instead, we focus on tools to enhance developer productivity and collaborate with technology partners like Google and Microsoft to optimize our tech stack.