Scaling Multilingual GenAI with Ishween Kaur скачать в хорошем качестве

Scaling Multilingual GenAI with Ishween Kaur

In this episode of In the Arena, host Allyson Klein welcomes Ishween Kaur, Generative AI Lead at Salesforce, for a practical conversation about what it really takes to deploy generative AI in enterprise environments. Ishween shares insights from her background in distributed systems and AI platform engineering, emphasizing the need to shift from a benchmark-first mindset to a real-world-first approach. She outlines actionable advice for practitioners navigating multilingual and multicultural GenAI implementations: start with one language and one workflow, measure trust before performance, add observability early, involve real users quickly, and design feedback loops with the assumption that your first version will be wrong. The discussion also explores how teams can move faster by iterating in production rather than waiting for perfect models or datasets, and how software engineers can thoughtfully transition into AI engineering. Host: Allyson Klein Guest: Ishween Kaur, Generative AI Lead, Salesforce Chapters 00:05 Welcome to In the Arena 00:12 Introducing Ishween Kaur 00:30 Ishween’s background in AI and distributed systems 16:45 From benchmark-first to real-world-first AI 17:12 Practical advice for multilingual GenAI adoption 17:25 Start with one language, measure trust first 18:09 Why iteration beats waiting for perfect models 18:57 Where to follow Ishween’s work 19:17 Building practical guidance for AI engineers 19:57 Closing remarks #generativeai #enterpriseai #aiengineering #techarena