From AI Experiments to Organizational Shift: Lessons from Mercari’s Transformation

Michael Galloway

VP, Platform Engineering @ Mercari

Snehal Shinde

Director, Platform Engineering @ Mercari

Abstract

Over the past year, Mercari has undergone an aggressive AI transformation across engineering workflows — from prototypes and code generation to AI co-pilots and multi-agent development systems. Through internal initiatives and a dedicated taskforce, we've learned that AI adoption isn't just a tooling challenge; it's an operating model challenge that requires fundamental shifts in how we work and think.

In this talk, we'll share real data and lessons from our journey: what worked and what didn't, why measuring AI impact is harder than expected, how we track behavioral, workflow, and system-level change, and how we're evolving toward an AI-first PDLC with guardrails, not gates.

This session is for leaders who want real lessons from the trenches of AI transformation — not hype, but practical strategy backed by data from a year of intensive AI adoption at scale.

About the speakers

Michael Galloway

VP, Platform Engineering

VP, Platform Engineering Michael Galloway is VP of Platform Engineering at Mercari, leading developer productivity, cloud infrastructure, and common services for Japan’s largest marketplace, with prior platform leadership at Netflix and HashiCorp. His current focus is exploring how agents can operate across the full development cycle, and what platform needs to become to support them.

Snehal Shinde

Director, Platform Engineering

Snehal Shinde is a Director in Platform Engineering at Mercari, Japan's largest C2C marketplace, leading Cost & Performance Engineering. She operates at the intersection of infrastructure efficiency, performance optimization, and developer productivity. Her career spans gaming, healthcare, and e-commerce, where she has built and scaled global engineering organizations in close partnership with product leaders. Currently, her focus is on improving Developer Experience by reducing friction in building and operating AI agents within complex brownfield ecosystems.