Luke
Han

It's late January 2015. Chinese New Year is three days away. Somewhere in Shanghai, a team of five or six engineers has made an unusual pact with their leader: no holiday until Apache Kylin can be installed in a single command. The project had just joined the Apache Software Foundation three months earlier, and the community was furious - the software couldn't run outside eBay's internal infrastructure. Luke Han, product owner of the project that would become his life's work, wasn't going home until it could.

They made it. The installer shipped. And something shifted in the Apache Kylin community almost immediately - the complaints about broken installs dried up, replaced by substantive discussions about cube algorithms and query optimization. It's a small story with a long tail. That sprint, that decision, that refusal to cut corners for the sake of a holiday, is a window into how Han operates: find the friction, remove it, then make it look easy.

Today, Han leads Kyligence as Co-Founder and CEO from dual headquarters in San Jose, California and Shanghai, China. Kyligence has raised over $110 million in venture funding, counts UBS, McDonald's, L'OREAL, Porsche, and China Construction Bank as clients, and has evolved from an enterprise OLAP platform into something Han calls an "AI-powered metrics intelligence layer" - the connective tissue between raw organizational data and the humans who need to act on it.

Han's path to data infrastructure was indirect enough to be interesting. He studied mechanical engineering at Zhejiang University of Technology from 1998 to 2002 - the kind of background that trains you to think in systems and tolerances, even if the systems you end up building run on distributed compute rather than steel. After graduation, he moved through enterprise software consulting, becoming Technical Director at Power Excellence East China and then Chief Consultant at Actuate China, where he led the sales engineering team.

In late 2011, he joined eBay as a Staff BI Architect in the Business Intelligence Platform Team. He was building data warehouse architecture for one of the world's largest e-commerce platforms at a moment when "big data" was transitioning from buzzword to engineering problem. By September 2013, he had become eBay's Head of Big Data Products and was leading the development of what would become Apache Kylin - a distributed OLAP engine capable of delivering sub-second SQL responses against petabyte-scale datasets.

When eBay contributed Kylin to Apache in 2014, it became something rare: the first Apache Software Foundation Top Level Project developed in China. That designation matters more than the title suggests. The Apache Software Foundation's TLP status is awarded to projects with demonstrated governance, community health, and technical merit - it's a statement that the project belongs to the world, not its creator. Han and the Kylin team had built something that crossed that threshold. He now serves as VP of Apache Kylin within the foundation.

In 2016, Han and the core Kylin team stepped out to found Kyligence. The name is a portmanteau: Kylin + intelligence. The idea was to build an enterprise-grade platform on top of the open-source technology - adding the automation, governance, cloud-native architecture, and AI features that large enterprises require but open-source projects typically can't prioritize. The company quickly attracted enterprise clients who needed OLAP performance at scale without building and maintaining the infrastructure themselves.

In the 2021 interview following Kyligence's $70M Series D, Han articulated the shift he saw happening in enterprise analytics: "The data analytics market has changed from decision support systems to democratized analytics - an 'analytics civilization,' if you like." Decision support was the old model: a handful of analysts with specialized tools making reports for executives. What Han is building toward is different - a world where every business user can query organizational data directly, in natural language, without understanding SQL or waiting for a data team.

The technical backbone enabling that vision is what Kyligence calls a unified semantic layer - a governed, centralized catalog of metrics and business definitions that sits between raw data sources and whatever interface users prefer: BI tools like Tableau and Power BI, Excel, or increasingly, AI-powered chat interfaces. The platform works with SQL and MDX simultaneously, which means it slots into existing toolchains rather than requiring organizations to rip and replace their BI investments.

One of Han's signature moves in positioning Kyligence has been the emphasis on Microsoft Excel as the "secret weapon" for data analysts. In a market full of startups trying to replace Excel, Han's bet is different: keep Excel, but let it query petabyte-scale datasets in real time through Kyligence's analytics engine. It's a pragmatic argument - one that speaks to the hundreds of millions of knowledge workers who already live in spreadsheets and aren't going anywhere.

In 2023, Kyligence launched its AI Copilot - Kyligence Copilot - built on Azure OpenAI. The product lets users query business metrics in natural language, receive automated root-cause analysis, get AI-generated business summaries, and even auto-create dashboards through conversation. Han describes it as AI augmenting human decision-making rather than replacing it. The platform analyzes KPIs autonomously, surfaces anomalies, and guides users toward root causes - functioning, in Han's framing, like "a domain expert available at all times."

Kyligence's customer roster spans industries that handle data at scale under regulatory pressure: UBS (financial services), China Construction Bank, MetLife (insurance), alongside consumer brands like McDonald's, L'OREAL, and Porsche. It's a client list that reflects Han's core argument - that real-time analytics at petabyte scale isn't just a tech problem, it's a business necessity across every sector where decisions depend on data.

Han holds an unusual combination of titles. As CEO of Kyligence, he's running a 120-person company with offices in San Jose, Shanghai, Beijing, Shenzhen, Wuhan, and New York. As VP of Apache Kylin within the Apache Software Foundation, he's stewarding the open-source project that underpins much of Kyligence's technology - maintaining community relationships, governance, and the project's technical direction. As a Microsoft Regional Director and MVP, he's embedded in Microsoft's enterprise ecosystem, which helps explain the Azure OpenAI partnership behind Kyligence Copilot and the Excel integration strategy.

The Fortune China 40 Under 40 recognition placed Han among the business leaders shaping China's technology industry. Solutions Review named him one of the "Coolest Data Analytics and Business Intelligence CEOs of 2021." In March 2023, he joined the Forbes Technology Council, an invitation-only community for world-class CIOs, CTOs, and technology executives.

As a speaker, Han has appeared at Strata Data Conference, Hadoop Summit, QCon, and ApacheCon across San Francisco, London, Beijing, and Shanghai. His presentations span both the deeply technical (cube optimization algorithms, distributed query planning) and the strategic (the future of self-service analytics, the transition from dashboards to AI copilots). He also contributes articles to Unite.AI and has authored public presentations on SlideShare going back to at least 2011.

Han's engineering philosophy shows up in a line from a 2021 interview: "A small investment in data storage can save a fortune in processing time." It's the core insight behind OLAP cubing - that pre-computing aggregated results costs disk space but returns query speed. Storage is cheap. Compute is expensive. People's time is more expensive still. Build for the expensive resource.

That logic extends to his product thinking. Kyligence's approach to analytics isn't to give users more powerful tools and trust them to figure it out. It's to pre-optimize, auto-tune, and surface recommendations so that users spend their time on decisions rather than data wrangling. The AI layer doesn't just answer questions - it asks them first, identifying which KPIs are behaving unusually before any human notices.

There's also something quietly principled about Han's continued commitment to Apache Kylin even as Kyligence competes commercially in the same space. Open source and enterprise commercial don't have to be adversarial. The open-source project builds community and establishes trust; the enterprise product delivers the production-grade reliability, support, and cloud-native features that large organizations require. Han has navigated that balance for nearly a decade without abandoning either side of the equation.