As 2024 approaches, industries have begun to shift their focus from conversations about generative AI and LLMs to building agentic AI frameworks for their businesses. People are even debating whether a single founder with a bunch of AI agents can run a company. This has also raised the question of the relevance of data scientists.
While you’re talking to GOAL, Indrajit Mitradirector of data science at Tredence, underlined the fact that agentic AI will dramatically disrupt industries and create great value. But instead of making data scientists obsolete, it will reshape their roles, skills and responsibilities.
Agentic AI requires a change in mindset and skills. Traditionally, data scientists focus on predefined problems: extracting insights and building models within clear problem frameworks. However, Indrajit noted that agentic AI needs data scientists to proactively identify complex problems and explore innovative solutions.
“The most important change is that data scientists need to frame problems, not just solve them. They must see themselves first and foremost as agents of business and understand the critical challenges businesses face,” Indrajit said.
Upskilling in the age of AI
To excel in this era, data scientists must develop a deeper understanding of business nuances and technical environments. While fundamental knowledge in statistics, machine learning and deep learning remains essential, the focus will shift to reinforcement learning, unsupervised learning and deep AI frameworks.
“Data scientists need to refocus their technical skills and, in turn, upskill themselves. They must develop expertise in agentic AI frameworks and platforms while mastering systems that integrate business insights and technical capabilities,” Indrajit added.
Furthermore, data scientists will no longer operate in silos. A strong understanding of broader ecosystems – cloud computing, DevOps practices and API integrations – will become critical. The ability to fine-tune the performance of multiple data sources and domains will be essential to delivering efficient and autonomous systems.
Data scientists as orchestrators in an agentic AI world
In a world where AI promises autonomous decision-making, many wonder whether these systems can function without data scientists. Indrajit is convinced that this is not possible. While agentic AI can function autonomously in specific contexts, data scientists remain central to designing, deploying, and optimizing these systems.
“Agentic AI cannot survive without data scientists. They are needed to design the solutions, train models, integrate systems and continuously monitor performance to meet business expectations,” Indrajit explains.
He used the analogy of a conductor in an orchestra to describe the evolving role of data scientists. Like conductors who understand the audience, the instruments, and the musicians, data scientists will orchestrate agentic AI systems to align business goals with technical execution.
“Data scientists will play the role of a master coordinator – a link between AI platform specialists, agentic AI frameworks and business stakeholders. Their success will depend on balancing these elements while ensuring seamless integration and efficiency,” Indrajit explains.
Ethics, governance and AI engineering
With the rise of agentic AI, ethical considerations, governance and responsible AI engineering become even more important. While these trends have already begun in industries such as healthcare, finance, and autonomous vehicles, their importance will only increase in the age of agentic AI.
Indrajit pointed out how AI is transforming industries such as healthcare, where AI-based diagnosis and patient management raise concerns about privacy, bias and transparency. Financial institutions are also integrating AI governance to comply with ethical and regulatory standards, such as the EU AI Act and the Dodd-Frank Act.
“Organizations are hiring data scientists with expertise in AI ethics to ensure responsible development of AI models. Data scientists will need to work with ethicists, regulators and legal experts to ensure that agentic AI systems are transparent, accountable and aligned with societal values,” Indrajit emphasizes.
The role of data scientists in multimodal AI
While agentic AI is one shift, the ever-growing adoption of multimodal AI presents another level of challenge. Multimodal AI uses different data inputs from a computer, such as text, images and audio, and independently generates insights. This has given rise to the idea that data scientists may be losing control of these systems.
Indrajit rejected this idea and emphasized that data scientists are best placed to overcome the challenges of multimodal AI. Their expertise is essential for ensuring the transparency, provenance and interpretability of data.
“Data scientists are critical to interpreting multimodal AI outputs and ensuring insights are delivered. They validate the authenticity of data, trace the input back to the source data, and continuously monitor the data. Techniques like attention mechanisms and saliency maps require human supervision, and data scientists are best suited for these tasks,” Indrajit further said.
The data scientist in the loop
The advent of agentic AI and multimodal systems marks a transformative phase for data science. Rather than replacing data scientists, these developments will expand their role and place them at the intersection of business strategy, technical innovation and ethical governance.
“Data scientists will play a critical role in translating the potential of agentic AI into real business value. They will act as orchestrators, balancing technical frameworks, business objectives and ethical considerations,” concludes Indrajit.
In this evolving landscape, data scientists must embrace new skills, deepen their domain expertise, and position themselves as indispensable leaders in an AI-driven future. By doing this, they will ensure that agentic AI systems are not only effective, but also fit business and societal needs.