Highlights:

• LLMs: The new frontier in NLP
• Challenges in deploying LLMs in real-world applications
• Overcoming obstacles to build robust LLM-powered solutions
• Insights from real-world LLM implementation

Abstract: Large Language Models, starting from BERT, and leading up to models such as ChatGPT, LLama, and Mistral, have revolutionized the field of Natural Language Processing. These models are now the go-to tools for any conceivable application dealing with language data. However, despite the promise shown by these models, there are significant practical constraints (such as cost, latency, accuracy, and privacy) that one needs to address while building real-world solutions using LLMs. In this talk, I will briefly talk about my experience of using LLMs in an industrial setting, the problems encountered, and the engineering problems we solved to make enterprise-grade solutions using LLMs.