AI & Data Product Manager
I build AI-powered tools that turn complex regulatory and compliance challenges into scalable, automated operations. 11+ years at Google and Roblox across product management, compliance, and Trust & Safety.
I came from pure humanities — linguistics, East Asian literature, translation — and taught myself to code through curiosity and necessity. I didn't set out to become a PM; I became one by building: automation scripts, SQL pipelines, AI systems, Chrome extensions. Eventually, I had built enough to prove I already was a PM.
What I do
RAG chatbots, bulk LLM inference pipelines, agentic systems, and AI-powered automation for enterprise operations at scale.
EU DSA, CMA Commitments, content moderation policy. Building systems that make legal obligations scalable and audit-ready.
Advanced SQL, Python pipelines, and end-to-end automation — replacing high-stakes manual processes with reliable, repeatable systems.
Translating complex regulatory and business requirements into technical direction across Engineering, Legal, Policy, and executive stakeholders.
Featured Work
Leading the design of Roblox's first-ever public API for researchers — a capability the company has never offered before. Enables academics, civil society organizations, and platform safety researchers to programmatically query Roblox's public data, advancing platform transparency in an entirely new way.
Designed an end-to-end RAG knowledge assistant for the 900-person Privacy Sandbox Working Group, indexing 12,000+ restricted documents. Won 2nd place in the Google Privacy Sandbox TPgM AI Hackathon.
Built a bulk LLM inference pipeline using SQLMiner to auto-summarize thousands of legal removal requests, surfacing AI-generated summaries directly in the team's primary management dashboard.
Built browser automation tools — a node.js proof-of-concept and a production Chrome Extension — that processed tens of thousands of cases, saving $50k+ in vendor time and eliminating a massive backlog.
Automated CMA-mandated quarterly reports with an E2E ETL pipeline — cutting generation time from 60–80 person-hours to under 5 (10x reduction), with zero errors and >95% auditability. "That was the smoothest cycle ever." — Legal
Built an interactive analysis dashboard over 10 years of Google's government content-removal data — 160 countries, 42 products, 22 removal reasons, 13 reporting periods. Turns a flat period-by-period report into a multi-year trend explorer with live filtering by country, product, and reason. Personal project built entirely in the browser.
View dashboard →Built a comparative analysis dashboard over EU Digital Services Act transparency reports from six platforms — Google (6 services), X, TikTok, Facebook, Instagram, Pinterest, and AliExpress — combining twelve separate reports into a single interactive view. Covers notices, proactive enforcement, government orders, and appeals for H2 2025. Personal project built entirely in the browser.
View dashboard →Skills
Publications & Awards
Other Work
Research Papers
Data Tools
Homework Help
Not to be confused with other Kieran Maynards in England. He is not me, and I am not he.