Shruti Pandey, MS 2016, leverages her data science skills to protect vulnerable children and families.

“Every number I look at represents a child, or a family,” says Shruti Pandey, Class of 2016, who is Director of Predictive Analytics at New York City’s Administration for Children’s Services (ACS). “I know that my work is directly changing lives, changing a family’s outcome. It’s very powerful.”

At ACS—which receives more than 65,000 calls each year to the city’s child protection hotline and manages roughly 50,000 active child welfare cases—Pandey leads a team using AI and predictive models to help caseworkers intervene earlier and prevent harm. Their tools analyze dozens of risk factors—from prior reports to housing instability–to flag high-risk cases for support. 

Turning Data Into Harm Prevention

One model her team developed identifies children at highest risk of severe harm or repeat maltreatment. Another predicts which families at highest are most likely to become unhoused, allowing the agency to prioritize them for housing vouchers before a crisis hits. As in all effective data science projects, the technical work is only as strong as the collaboration and communication behind it. “We start by partnering with agency divisions to define the real-world problems we’re trying to solve and clarify what needs to be predicted or optimized,” Pandey explains. 

Her team works closely with groups like Child Protection and Preventive Services to identify where predictive analytics and data-informed decision making can improve workflows and outcomes. 

“From there, we develop a strategy, build and test models, and refine them iteratively—adjusting parameters, weighing competing priorities, deciding how wide or narrow we want our net to be and continuously evaluating the effectiveness of our approach.”

This technical process is always guided by a larger commitment: ensuring that the tools are not just effective, but also ethical.

GenAI has real potential to surface insights hidden in unstructured data, but it needs to be deployed thoughtfully, and with rigorous oversight.

– Shruti Pandey, Class of 2016
Data for Good in Action: Ethical Innovation in an Evolving AI Landscape

Pandey is also at the forefront of ethical AI governance in the public sector. In 2024, ACS became one of the first NYC agencies to formalize an AI governance policy, building on federal frameworks and grounding its work in transparency, bias mitigation, and privacy safeguards. 

Now, generative AI tools are under consideration for integration into the workstream for tasks like analyzing a vast corpus of caseworker notes. 

“Creating the right model is going to be a challenge,” Pandey says. “GenAI has real potential to surface insights hidden in unstructured data, but it needs to be deployed thoughtfully, and with rigorous oversight.” 

A Public Sector Path Forged at DSI

Pandey began her journey into public service as a student at the Institute. A summer internship with New York City’s Fire Department underscored the power of data to solve real-world problems. 

“One of my first projects was during an outbreak of Legionnaires disease, which spreads through a bacteria in water. Working with the Mayor’s Office of Data Analytics , we built a model to predict which buildings had potentially dangerous cooling towers. The fire department used that model to inspect and intervene,” she recalls. “That demonstrated how data science could directly improve public health and safety. I knew I wanted to stay in the public sector.” Now, she hosts Institute interns at ACS each summer, giving them the same kind of hands-on experience that shaped her own career.

A Message to Future Data Scientists

Beyond the direct impact that first drew her to the role, Pandey says public service has opened doors she didn’t expect—putting her in conversation with researchers across the country and exposing her to new tools and ideas. “Working in AI governance enables me to develop and implement the “do’s and don’ts” that affect multiple organizations or agencies. I don’t think you can get that kind of experience in any other sector.”

Whether she’s building models to prevent harm, shaping ethical AI policy, or mentoring future data scientists, Pandey’s career reflects the heart of DSI’s mission: using data for good.