Vanja Josifovski

The most useful business data — such as transactions, customers, and inventory — lives in your data warehouse. Yet AI models have historically struggled with relational structures, requiring time-consuming feature engineering or lossy workarounds to use it for predictive or generative AI applications. The Kumo platform uses breakthroughs in Graph Transformers and Relational Deep Learning (RDL) to let you build highly accurate predictive and embedding models directly on relational data with zero feature engineering. This lets data scientists and engineers ship more accurate predictive and generative AI applications across the business up to 20 times faster than with traditional methods.

Homes is the core business of Airbnb including the whole spectrum of accommodations from shared rooms and houses to hotels and real estate developments in all tiers: Regular, Plus, Lux. Leading an organization that has both horizontal groups as Homes Engineering and Homes Data Science as well as GM groups as Marketplace, Relevance and Personalization, and Regulatory Frameworks. All with the goal of making the world a smaller place for everyone!

As CTO, responsible for the overall technical vision and strategy of the company and communicating that to leadership and the teams; hiring and development of technical talent. As VP of Engineering managing some of the core eng teams.

Responsible for all aspects of the day-to-day operation, including revenue bottom line, operational stability, and team health. On the technical side, working with the team on a rewrite of the ads infra stack and evolving the quality metrics, goals, and modeling approach.

Leading multiple engineering organizations at the core of the Pinterest product: Search, Homefeed, Visual Search, Topics, etc. Our goal is to create the most relevant, personalized content to deliver a great user experience and drive engagement.

Responsible for the user growth company-wide metrics, which is the #1 priority for the company. Worked out a playbook for international growth. Grown the team from ~20 to ~70. Areas of focus: user acquisition (SEO), signup flow, activation, re-engagement with emails/notifications, collaboration, and virality. Turned metrics up after multiple weak quarters. Got Pinterest to grow internationally.

Back to writing code. co-wrote a runtime of a search engine leaf node. Deployed for serving ads

Mobile retargeting using AI. Acquired by Pinterest

Led a portfolio of projects in Google Research with a mix of fundamental ML research and using large-scale machine learning, search, and information extraction technologies to improve multiple Google products. Direct impact on user experience. Supported numerous internal customers, including search ads, youtube, google play, etc.

Teach Computational Advertising class at Stanford's MS&E Department in the Fall semester. Some of the class content: http://www.stanford.edu/class/msande239/

Leading a research team in a portfolio of advertising technology projects, including next-generation targeting platform, ad exchange design, and new targeting initiatives.

Worked across the stack on the DB2 relational database: query compiler, runtime, XML runtime, enterprise search