Skip to content
Sebastian.Gebhardt

Sebastian Gebhardt — Santiago, Chile

I believe Chile can produce world-class retail.

I’m building it through distinctive stores and AI systems that make the business better.

CEO of Yáneken · Co-founder of Menlo & Oak

Explore my AI work Read my story

My work brings together engineering, software and business judgment. I build from Chile, with the conviction that we can compete with the best in the world.

The conviction, built in retail

Yáneken · The operation where these ideas are put to work

stores
160+
brands
10+
people
1,500

Figures as of August 2026.

In the conversation
La TerceraPulsoDiario FinancieroIdDCMás Que Un PitchStanford GSB
Selected work

From problem to system.

Three ways I work with AI: simulating a decision, improving an operation and anticipating demand.

Marketing intelligence

Andrea

The simulation marketing was missing: a 100,000-person synthetic society, creative auditing and Bayesian attribution in one system.

In production · 8 brands

The decision behind the system

Bring research, creative evaluation and measurement into the same workflow. A simulation can help explore an idea; an auditor can help revise a piece; attribution over actual sales asks what happened after spending. Each answers a different question. The useful system connects those decisions.

Explore the case
100,000 synthetic respondents · a focus group convenes in seconds
System overview

Store operations

IRIS

Retail's operating layer, running on WhatsApp: photo-based visual validation, voice-note ticket routing, and AI-answered Google reviews.

In production · 160+ stores

The decision behind the system

Start with WhatsApp. Introducing another application would make adopting the interface a prerequisite for solving the operating problem. Photos and voice notes already travel through the stores’ working day; IRIS gives those inputs standards, routing and follow-through.

Explore the case

Demand forecasting

Delphine

Multi-modal sell-through prediction for footwear: a photo of the shoe plus its attributes, and a buy/pass verdict before the season.

Deployed · GCP

The decision behind the system

Combine the product image with structured attributes and price. The published model uses a late-fusion network with EfficientNet-B0 for photography. Its output reaches the buyer through Telegram, together with an uncertainty band. The final buying decision stays with the person.

Explore the case
See all nine systems
Writing

Start here.

A technical decision, a retail bet and the story of why I came back to Chile.

All writing
Newsletter

Subscribe to The Retail Brief.

Every other week, one question about retail, AI, and building from Chile. Available on LinkedIn and Substack.

Sebastian at Pixar Pier, California

Building is also choosing where.

I’m an engineer, third generation in a family business and Alexander’s dad. After Stanford, I came back to Chile to build. AI, retail and people are part of the same story.

More about me

What I’m working on

Yáneken's internal workshop — IRIS validating stores over WhatsApp across 160+ stores, Draper piloting commercial agents with Hoka, and Project Cerebro putting the whole market on the pricing table. What we used to buy as SaaS, we now build.

Visit my Now page
Contact

Tell me what you’re building.

A problem to solve with AI, a project to build together, or a conversation worth having. Here are the best ways to start.

hello@sebastiangebhardt.comBook a conversation