I help professionals translate complex business problems into scalable and trusted AI solutions. I advise leaders on AI strategy, which is an interplay of people, knowledge, data, processes and technology.
Pınar Kahraman
About
My background spans operations research, statistics, bioinformatics and PhD-level research. I have shipped automated decision making systems into production, led engineering teams, and owned AI products at global scale. Along the way, I have developed a deep understanding of the probabilistic nature of AI, what role data plays in decision making, as well as the importance of observability and monitoring.
This combination means that I frequently sit with AI engineers as well as leadership: translating between probabilistic reality and institutional decisions, and being candid about feasibility, readiness and risk.
Expertise
Currently building a multi-agent copilot that holds the customer's hand throughout their journey with the company: answering questions, surfacing guidance and taking actions on their behalf.
Led delivery of an enterprise agentic platform that reviews customer marketing campaigns across messaging channels: LLM-based text moderation, image recognition for visual validation, and human-in-the-loop escalation feeding continuous training. 95% automated review coverage, 60% faster reviews and roughly $300K annual operational savings.
Directed product strategy and stakeholder alignment for EMEA's centralised forecasting platform — standardised forecasting methodology, centralised data features, shared operational metrics and a scalable AI operating model — plus the information architecture for end-to-end supply chain analytics.
Built a scalable real-time anomaly detection service for the technology organisation, enabling proactive monitoring and early incident detection, and shaped the design of the Global ING Monitoring Platform together with IT architects.
Developed a Python forecasting web app predicting incoming payment transaction volumes with dynamic real-time adjustments — published as "Predicting the daily number of payment transactions in the largest bank in the Netherlands" (IEEE Big Data).
Built an optimisation model predicting the metabolic strategies micro-organisms adopt, aimed at industrial use — the work behind my instinct for constrained optimisation, and where I developed and taught Calculus for Systems Biologists.
Classified calcium channel antagonists with a hyperbox approach and a mixed integer nonlinear programming model for medical drug design, published in Industrial & Engineering Chemistry Research.
Experience
2025 — now
Global Senior Product Manager, AI
Leading AI product strategy: conversational analytics, data agents, and an agentic compliance platform with human oversight.
2023 — 2025
Lead AI Product Manager
Directed the foundational EMEA demand forecasting strategy and the information architecture for supply chain decision intelligence.
2016 — 2022
Data Scientist → Senior Product Manager
Built forecasting tools and led a ten-person engineering team developing real-time anomaly detection at the core of a global monitoring platform.
2014 — 2016
Advanced Analytics Consultant
Connected European analytics specialists and applied advanced modelling to industry challenges.
Teaching
Wherever I work, I end up building courses. I have lectured at Delft University of Technology and designed programmes that help engineers, analysts and leaders engage confidently with analytics and AI — without hand-waving about how the models actually behave.
TU Delft / edX
AI in Practice: AI Applications in FinTech
ING
Analytics for DevOps
Vrije Universiteit
Calculus for Systems Biology
Vrije Universiteit
Bioinformatics in Practice
Pro-bono advisory
Please do reach out if: