Sanjana Arun
Product Lead, eBay
San Francisco, California, United States
Actions
Sanjana Arun is a Product Lead at eBay, where she leads AI-powered Search and Advertising systems that determine how visibility, relevance, and monetization are allocated across large-scale marketplaces. Her work focuses on translating machine learning innovations in ranking, retrieval, and personalization into trusted, high-impact experiences for buyers and sellers.
She specializes in designing AI systems as decision engines, where models, data, and feedback loops shape user behavior and marketplace dynamics. Sanjana serves as an Ambassador for Product-Led Alliance leading panels, speaking and writing about AI. She is also an active speaker on Grace Hopper, Women In Tech, ProductWorld, AI Dev Summit among others. Her work has been published on Mind the Product and CMSWire websites.
Links
Area of Expertise
Topics
Debugging AI Systems in Production: A Hands-On Failure Analysis Workshop
AI systems in production rarely fail due to obvious errors in model performance or system availability. Instead, they degrade through more subtle mechanisms: objective misalignment, feedback loop amplification, distribution shift induced by the system itself, and unintended interactions between model outputs and downstream business logic. In many cases, core metrics remain stable while real-world outcomes deteriorate, making these failures difficult to detect using traditional monitoring approaches.
This hands-on workshop introduces a structured, system-level approach to debugging production AI systems.
Drawing from large-scale search, recommendation, and advertising systems, the session frames AI products as multi-stage decision pipelines composed of retrieval, ranking, allocation, and feedback layers. Participants will learn how failures propagate across these layers and how to trace issues beyond model predictions into system behavior over time.
The workshop focuses on diagnosing four common classes of production failures:
1) Objective function misalignment between model optimization and business outcomes
2) Feedback loop dynamics that amplify bias or distort system behavior
3) System-induced distribution shift, where model outputs alter future training data
4) Local optimization effects that degrade global system performance
Participants will work through real-world scenarios to:
- Decompose AI systems into their decision layers and identify where control and logic reside
- Trace metric movement across stages of the pipeline to isolate root causes
- Analyze how user behavior and system outputs interact to create emergent failure modes
- Propose mitigation strategies, including constraint design, multi-objective optimization, and controlled exploration
By the end of the session, attendees will leave with:
1) A practical framework for debugging AI systems beyond model accuracy and offline metrics
2) Techniques for tracing failures across the end-to-end AI decision stack
3) Methods to detect and reason about feedback loops and system-driven distribution shift
4) Approaches to aligning model objectives with real-world system outcomes
5) A reusable mental model for designing more observable, robust, and production-ready AI systems
OPEN Session: Algorithmic P&L: Who Owns the Roadmap When the Model Makes the Decisions?
AI is quietly rewriting the rules of product ownership.
As models shift from supporting decisions to actively shaping outcomes, the traditional software roadmap begins to break down. Plans become probabilistic. Performance may improve - but accountability becomes harder to define.
When outcomes drift, costs rise, or tradeoffs surface, leadership teams are left asking a new question: who actually owns the result when the model is making the decisions?
This session introduces the concept of Algorithmic P&L - a new way of thinking about ownership in AI-driven products where financial outcomes, user experience, and system behavior are increasingly determined by learning systems rather than deterministic plans.
Drawing on real-world patterns from large-scale AI platforms such as search, ads, and marketplaces, the talk examines how responsibilities across product, engineering, and data science collide - and why legacy ownership models no longer hold.
Rather than focusing on governance frameworks or org charts, this talk surfaces the leadership gap that emerges when models, not humans, effectively own the roadmap. Attendees will leave with a clearer mental model for assigning accountability, making decisions, and leading teams when outcomes are driven by algorithms rather than specifications.
What you'll takeaway from this session -
- Why AI-driven products create ownership and accountability gaps at scale
- How probabilistic systems break traditional roadmap and planning assumptions
- What leadership teams must rethink when models - not plans - shape outcomes
Upskilling in the Age of AI: What Every Product Manager Needs to Learn (and Unlearn)
AI is no longer a feature - it’s fast becoming the foundation of how modern products are built, tested, and scaled. For product managers, that shift demands more than new tools; it requires new ways of thinking. The skills that defined great PMs a decade ago - prioritization, storytelling, experimentation - now need to evolve alongside data literacy, agent collaboration, and model awareness.
In this session, Sanjana Arun, Product Lead at eBay, explores how PMs can future-proof their craft in the age of AI and intelligent agents. Drawing from her experience leading AI-driven Search and Ads products, she’ll share what to double down on, what to let go of, and how to stay relevant as AI reshapes every layer of product work - from discovery to delivery.
Key takeaways to expect from this session:
1) Learn a practical framework to assess your own AI readiness across core PM competencies
2) Discover how to pair traditional product instincts (like prioritization and user empathy) with emerging skills in data fluency and model interpretation
3) Build an actionable roadmap for integrating AI and agent tools into your daily workflow - from ideation to experimentation to launch
Please note that Sessionize is not responsible for the accuracy or validity of the data provided by speakers. If you suspect this profile to be fake or spam, please let us know.
Jump to top