Senior PM working the decision layer between LLMs and the enterprises that depend on them.

I treat AI-native products as probabilistic systems, not feature trees. Acceptance criteria, eval frameworks, and shared accountability with research replace the SaaS-era spec. The PM owns the “is this good enough yet?” question. Engineers and researchers own the “how do we get there?” question.

I prefer compound AI over single-model architectures for any decision that matters. LLMs handle language, semantic routing, and explanation. Deterministic logic, ML models, and search handle constraints, risk classification, and retrieval. The LLM should rarely be the decider.

I default to writing things down: problem statements, decision logs, eval results, fund-or-kill business cases. Tight pairing with applied research and ML engineering early; loose coupling with design once a v1 surface is locked. Scope cuts beat timeline extensions.

• AI-native enterprise SaaS, 0-to-1 to GA under end-to-end PM ownership
• Decision intelligence: orchestration, explainability, pre-production validation
• Data and AI/ML asset governance: catalog, lineage, classifiers, knowledge graph
• LLM integration in production: provider strategy, eval frameworks, compound AI
• Microsoft co-sell and Marketplace go-to-market for ISV products
• Regulated-industry product strategy: financial services, insurance, healthcare, public sector

• 2024 to NowSAS · Senior Product ManagerSAS Decision Builder. Concept to Microsoft Marketplace GA in under twelve months. Owned product, GTM, partner enablement, and Microsoft co-sell end-to-end.
• 2021 to 2024SAS · Senior Product ManagerSAS Information Governance & SAS Lineage. 0-to-1 platform for data and AI/ML asset governance. GA in 2021, scaled to 100+ deployed enterprise customers.
• 2018 to 2020SAS · Principal Technical ArchitectPre-sales solutions architect across the SAS data management portfolio. Global 500 reference architectures and POC design.
• 2012 to 2017SAS · Senior Software Engineer / Tech LeadSAS Data Preparation and Profiling. The upstream surface for analytics and ML on enterprise data.

• MBABoston University Questrom
• MS CSIllinois Institute of Technology
• BE CSBharathiar University, India

• Real decision authority on what the product becomes, not just what ships.
• AI-native scope where the AI is the substrate of the product, not a deliverable bolted on.
• A team that takes evals and decision validation seriously.
• Leadership that understands probabilistic systems and their regulated-industry implications.
• Customers who feel the impact in decisions and dollars, not dashboards.