At the world’s largest retailers, a product roadmap is rarely just a backlog of features. It is a live bet on how millions of customers will shop next quarter, how associates will run their shifts tomorrow, and how billions of dollars of inventory will move through a supply chain that never truly sleeps.
For Sandeep Mahajan, Director of Data Science Product Management at Walmart, this reality has defined more than two decades of product leadership across some of retail’s most recognized organizations, including Sears Holdings, Amazon, Levi Strauss & Co., and Walmart. His work has spanned large-scale RFID deployments, merchandising automation, enterprise data platform modernization, and more recently, agentic AI systems designed to orchestrate store operations in real time.
While technologies have evolved dramatically, Sandeep observes that the core challenge of retail product leadership has remained constant: transforming ambiguous, cross-functional problems into simple, scalable systems that create value simultaneously for customers, associates, and the business.
The arc of his career reflects how retail innovation itself has evolved, from working within constraints, to building scalable mechanisms, to creating reusable platforms, and now to enabling agentic AI systems.
Learning to Work Within Constraints: The Sears Years
Sandeep’s early immersion into large-scale retail began at Sears, an environment defined by tight margins and deeply entrenched legacy systems.
What initially appeared to be a straightforward mandate, improving margins and reducing aged inventory, quickly revealed itself as a complex organizational challenge spanning merchandising, supply chain, stores, and finance.
He co-led a chain-wide initiative that reduced aged inventory exposure by approximately $100 million, requiring coordinated changes in fulfillment logic, promotional strategy, and store execution across hundreds of locations.
One defining moment came during an RFID rollout across men’s apparel departments in more than 200 stores. A store leader initially questioned the accuracy of RFID-generated inventory counts, preferring manual verification. Walking the sales floor together and reconciling discrepancies item by item helped demonstrate the system’s reliability. Adoption accelerated soon after, reinforcing a lesson that would shape Sandeep ’s leadership philosophy: technology succeeds only when frontline teams trust it.
Through these experiences, Sandeep developed a foundational belief that in mature retail environments, constraints are not obstacles but design parameters. Effective product leadership requires navigating legacy realities while still delivering measurable outcomes.
Designing for Scale: Automation at Amazon
Sandeep ’s transition to Amazon introduced a distinctly different operating philosophy centered on scalable mechanisms and automation.
As a Senior Product Manager supporting Amazon Books’ physical retail expansion, he focused on automating merchandising workflows that previously required significant manual coordination. Category teams spent extensive time building assortments, managing marketing assets, and synchronizing updates across locations.
Observing category managers manually managing spreadsheets ahead of store launches highlighted a critical limitation: expansion could not scale if operations depended on manual processes.
Sandeep helped translate tacit merchandising expertise into system-driven rules for assortment planning and placement, automated marketing asset distribution, and established feedback loops connecting store performance data directly into decision models.
The initiatives delivered roughly 90% time savings for category teams. More importantly, they redefined managerial span of control, enabling new physical retail formats to scale without proportional increases in headcount.
This period reinforced a principle that continues to guide his work: lasting product impact comes less from individual features and more from repeatable mechanisms that operate reliably across teams and time.
Making Data Actionable: Global RFID at Levi Strauss & Co.
At Levi Strauss & Co., Sandeep focused on scaling RFID capabilities across a global retail footprint spanning the United States, Europe, and China.
The challenge was no longer adoption but operationalization, transforming RFID from a technology deployment into a durable enterprise capability.
Three priorities guided the effort:
Inventory Accuracy as a Business Metric
RFID outcomes were tied directly to on-floor availability and measurable sales improvements rather than technical performance metrics.
Standardized Global Playbooks
Regional variations were acknowledged while deployment frameworks remained consistent enough to enable repeatability at scale.
Actionable Interfaces
RFID insights were embedded directly into workflows used by merchants and store teams, ensuring data translated into decisions rather than remaining confined to dashboards.
This phase of Sandeep ’s work reinforced a recurring insight: enterprise value emerges when initiatives evolve into reusable platform capabilities that can be redeployed across brands, regions, and channels.
Rewiring the Retail Core: Walmart’s Platform and Store AI Journey
At Walmart, Sandeep ’s work expanded into one of the most complex retail ecosystems in the world, where improvements must demonstrate measurable impact at enterprise scale.
His early focus centered on merchandising platforms and item master data modernization. Legacy systems were not designed for an omnichannel retail environment in which products exist across formats, channels, and supplier ecosystems.
The initiative introduced a modern master data management platform that significantly expanded item attribution and enabled structured, supplier-led onboarding. The outcome improved assortment availability, strengthened analytics accuracy, and established a unified source of truth across downstream systems.
More recently, Sandeep has led initiatives within Agentic Intelligence and Store AI, focused on building intelligent store agents capable of:
- Prioritizing operational tasks across safety, freshness, modular execution, and asset protection
- Interpreting real-time store conditions using data, devices, and computer vision
- Guiding associates through daily workflows while preserving human judgment
The Store AI portfolio represents a frontier in how large retailers operationalize agentic AI, moving from isolated pilots to production systems influencing how work is prioritized across thousands of locations.
During an early pilot, a fresh department manager remarked that the system felt like it was “looking ahead” rather than simply recording completed tasks, a signal that the technology was reshaping how work was experienced on the store floor.
In this space, product leadership combines AI strategy, organizational change management, and deep empathy for frontline operations.
An Evolving Philosophy: From Constraints to Agentic Systems
Across Sears, Amazon, Levi’s, and Walmart, a consistent progression emerges in Sandeep’s product philosophy:
- Treat constraints as design inputs.
- Build mechanisms instead of one-off solutions.
- Transform mechanisms into reusable platform capabilities.
- Use those capabilities as the foundation for agentic systems that orchestrate work in real time.
Technology has changed dramatically, but the evolution of ideas has followed a clear and continuous path.
What Hasn’t Changed in 20 Years
Despite rapid advances in AI and data, Sandeep identifies several enduring constants in effective retail product leadership:
Product leaders act as translators and integrators.
The most valuable problems exist at organizational boundaries, between technology and merchandising, headquarters and stores, finance and operations.
Impact outweighs elegance.
Solutions that scale across hundreds of locations and deliver measurable outcomes outperform theoretically perfect systems that remain confined to pilots.
People and processes determine success.
Technologies deliver value only when employees understand and embrace new ways of working.
Balancing long-term vision with short-term delivery is essential.
Retail demands quarterly outcomes even as foundational platform investments unfold over years.
The Future of Retail Product Leadership
Looking ahead, the scope of retail product leadership continues to expand.
AI is evolving from copilots into autonomous agents capable of decision-making and execution. Data is becoming increasingly real-time and operationally embedded, while expectations around governance and transparency continue to rise.
Yet the central responsibility remains unchanged: understanding systems rather than isolated features.
Whether modernizing data platforms, scaling global RFID capabilities, or orchestrating agentic AI in stores, product leaders must continuously ask:
- What real problem is being solved for customers and associates?
- How will the solution scale across regions and business cycles?
- How will success be measured and adapted over time?
After more than two decades in retail technology and product leadership, Sandeep remains optimistic. Advances in AI and data are creating unprecedented opportunities to design systems that make retail simultaneously more efficient and more human-centered, a direction that continues to define the industry’s next era.
Author Bio
Sandeep Mahajan is Director of Data Science Product Management at Walmart, where he leads Agentic Intelligence and Store AI initiatives. He has previously held product leadership roles at Sears, Amazon, and Levi Strauss & Co., focusing on large-scale data platforms, AI systems, and retail innovation.
