The AI tech my dad helped pioneer is now the foundation for the tools I build at AT&T

The advent of advanced AI tools is reshaping industries, yet they still have limitations. Natalie Gilbert, a 30-year-old data scientist at AT&T, reflects on how her father's pioneering work in speech recognition laid the groundwork for her current projects in generative AI. Her father, Mazin Gilbert, contributed to groundbreaking research at Bell Labs, collaborating with luminaries like Yann LeCun and Dennis Ritchie. Growing up in her father's workspace, Natalie often observed intense discussions and complex diagrams being crafted on whiteboards. This environment sparked her creativity and analytical thinking, leading her to create her own decision trees. One memorable project they worked on together was Dr Bot, an early large language model designed to provide symptom assessments and care directions. Natalie's current role involves developing AI agents that utilize decision-making frameworks to streamline processes, such as identifying relevant HR policies within AT&T's vast organizational structure. This innovative approach aims to simplify how employees address HR issues, showcasing the transformative potential of AI in large corporations. While Natalie employs a coding assistant to enhance her efficiency, she emphasizes the necessity of understanding the underlying technologies behind AI, particularly as digital assistants become more prevalent. A common pitfall for developers is neglecting the foundational knowledge of machine learning models, which can lead to ineffective AI solutions. In a rapidly evolving landscape, Natalie notes the shift towards natural language processing over traditional coding. Her work now heavily revolves around prompt engineering—crafting effective queries for machines to comprehend human intent. This evolution, she finds, echoes her father's contributions from decades ago, reinforcing her connection to his legacy as she navigates the future of AI.