The role of AI and LLMs in data — Pedram Navid // Dagster Labs

Pedram Navid, Head of Data Engineering and DevRel at Dagster Labs, explores data products and AI and large language models’ roles in data. While AI has limitations, it offers data practitioners new ways to explore and leverage data, transforming the analysis process. Further, AI and LLMs democratize access to data analysis for non-technical people, enabling them to explore and derive insights from complex datasets without feeling overwhelmed. Today, Pedram discusses the role of AI and large language models in data.
About the speaker

Pedram Navid

Dagster Labs

- Dagster Labs

Pedram is Head of Data Engineering and DevRel at Dagster Labs

Show Notes

  • 01:29
    The role of AI in data teams and its limitations
    AI, particularly models like GPT, provides a companion for data practitioners, helping with creative problem-solving. While AI can automate certain data tasks, its limited in its tribal data knowledge, making it more of an assistant than a replacement for data professionals.
  • 06:05
    AI's ability to enhance data analysis
    AI can uncover new connections and perspectives in vast datasets. However, AI provides data practitioners with novel ways of exploring data and analysis, and the ease of conversing with a bot provides an intermediary between reading a book and random Google searches.
  • 08:44
    The impact of AI on casual data users
    AI and LLMs make it easier for non-technical users to work with data without feeling overwhelmed or needing to bother data experts. The natural language processing ability of these tools enables individuals to interact with and derive insights from datasets using everyday language.
  • 10:53
    Empowering non technical users with AI
    AI tools reduce the friction to get started using data to solve problems. While some hesitation remains, the newfound ability to explore and analyze data independently fosters a sense of control and reduces reliance on human experts for routine queries.
  • 13:12
    Mitigating privacy risks with AI
    Consider your organization's risk policies and opt for AI tools with enterprise-grade assurances to prevent privacy lapses. For personal use, exercise prudence, avoid sharing sensitive information, and explore specialized tools designed to address privacy and security concerns.
  • 14:42
    Strategic considerations for AI integration
    Developing an AI strategy within an organization is crucial for staying competitive and closing the productivity gap. While AI is valuable for productivity, avoid over-reliance on it to maintain critical thinking and problem-solving abilities.
  • 17:29
    Leveraging AI evangelists to drive adoption within organizations
    Large companies can benefit from identifying employees who can act as AI advocates to encourage adoption internally. Enabling passionate and motivated employees to drive adoption keeps them motivated while creating a culture of innovation.

Quotes

  • "AI has become more of an assistant than a replacement for data people. It's been a rubber duck you can talk to. Will that change in the future? Maybe. But I'm less optimistic on that front." - Pedram Navid

  • "The real unlock with LLMs and AI is that it becomes easier for everyone to become a data person. They enable people who are less technically capable or less experienced in data to feel more comfortable operating on data." - Pedram Navid

  • "When it comes to AI and LLMs, be sensible and reasonable. Don't put something in that you wouldn't want others to see and know about in these open channels." - Pedram Navid

  • "You need an AI strategy. If you don't, your competitors will have one, and a productivity gap is going to occur." - Pedram Navid

  • "Even if youre a skeptic, play around with AI. Its better to understand it now than four years from now when it takes over the world." - Pedram Navid

About the speaker

Pedram Navid

Dagster Labs

- Dagster Labs

Pedram is Head of Data Engineering and DevRel at Dagster Labs

Up Next: