Career
I'm an engineering leader and builder, currently VP of Engineering for AI products at Figma. Over the past 20+ years, I've helped build and scale products at Dropbox, Uber, and LinkedIn. I care about developing strong teams, solving hard technical problems, and turning emerging technology into products people find useful.
My work spans consumer products, distributed systems, machine learning, and AI. Below are a few highlights; see my LinkedIn profile for more details.
You can also find my articles, interviews, talks, and presentations around the web.
Figma
2026–present
I lead engineering for Figma's AI products, helping build the future of design. The work brings together model training, evaluation, multimodal research, agents, and product engineering, with a focus on the craft that makes great design tools useful.
In How we’re training a Figma-fluent editing model, I shared our approach to training Figma's first-party editing model, including supervised fine-tuning, reinforcement learning, and design-specific evaluations, and how it works alongside frontier models in the Figma agent.
Dropbox
2024–2026
I was head of engineering for Dropbox's AI platform and Dropbox Dash, the AI assistant and context engine for work. My organization spanned AI research, platform infrastructure, and the Dash product experiences across web, mobile, and desktop.
- AI research and platform: Responsible for research, evaluations, inference, training, multimodal capabilities, model routing, agent harnesses, memory, tool use, and context engineering.
- Data, search, and context: Responsible for data infrastructure, connectors, the context graph, search and discovery, and lexical and vector retrieval.
- Dash product experiences: Led engineering for chat, agents, stacks, and search across web, mobile, and desktop.
- Engineering leadership: Served as engineering lead for three M&A deals: Jamscape, Promoted.ai, and Mobius Labs.
I wrote about this work in knowledge graphs, MCP, and DSPy in Dash and Building Dash: RAG, multi-step AI agents, and business users.
Uber
2015–2023
For most of my time at Uber, I led engineering for the core Uber Eats product. During my tenure, Eats served over 7 billion orders across more than 30 countries. As part of Uber's engineering executive team, I also helped guide systems scaling, operational health, tooling, and engineering culture across an organization of 4,500 engineers.
- Core Eats experience: Led engineering across the consumer product, including rewrites of ubereats.com, the iOS and Android apps, and the merchant tablet app.
- Postmates integration: Led the engineering integration effort for Uber's $2.7B acquisition of Postmates, involving more than 400 engineers across the company.
- Machine learning: Built the Eats machine learning and recommendations team to improve food discovery, home feed recommendations, and search.
- New ways to order: Initiated the Eats Pickup business and developed the cross-platform strategy that brought Eats ordering into the main Uber app.
- Driver acquisition: Built the engineering organization that helped onboard millions of drivers to the Uber marketplace.
Selected press and engineering stories from our team's work
LinkedIn
2011–2015
At LinkedIn, I led teams building member profiles, education products, and mobile experiences during a period of rapid growth.
- Profile 2.0: Led the rebuild of LinkedIn's profile page and frontend stack, coordinating 35 core engineers and more than 100 engineers across the company.
- Project Voyager: Helped lead the flagship mobile app rewrite, presenting progress to the board and CEO, guiding API and iOS architecture, and leading the My Network product team.
- Higher Education: Built and led a new organization serving students and universities, including University Pages, alumni tools, and school rankings.
- Infrastructure: Co-owned and scaled the member profile database to more than 500,000 queries per second.
Lockheed Martin
2005–2011
I started my career building sensor and data fusion systems at Lockheed Martin. I led the Distributed Fusion Manager team, working on fault-tolerant distributed systems over mobile networks, and contributed to tracking algorithms, Kalman filters, and bias correction. I was also selected for the three-year Engineering Leadership Development Program.
Still building
I continue to build my own products. Working through the details of code, design, and customer feedback keeps me close to the craft.
- Draft Punk — My fantasy football draft companion, combining custom rankings, mock drafts, and recommendations. I've been building it since 2010.
- Earthquake Alert! — An app for exploring earthquakes, maps, and news.
- AI search experiments — Hands-on exploration of retrieval and language models, including Yaddle and a guide to building an AI search app.
Selected writing
- Building Dash: RAG, multi-step AI agents, and business users — Turning AI systems into useful workplace products.
- A Brief History of Scaling Uber — The engineering decisions behind global growth.
- A Brief History of Scaling LinkedIn — Scaling systems and engineering culture. It was even translated into Chinese. The discussion of Project InVersion was referenced in The DevOps Handbook and Wiring the Winning Organization.