Build the performance layer for the agentic shopping era.
We are a small team building Lantern in New York. The work sits at the intersection of model evaluation, commerce systems, and design for serious operators.
Three threads of work, one operating loop.
Each thread compounds. The systems team makes the answers observable; the commerce team turns signals into shippable work; the measurement team proves what moved.
Make AI answers observable
Run the questions shoppers ask across ChatGPT, Gemini, Perplexity, DeepSeek, Grok and Alexa, and explain why a recommendation moved.
Translate signals into commerce work
Turn what those answers show into the product data fixes and content a store team can approve and ship this week.
Prove what moved
Tie each shipped fix to the AI visits, orders and revenue that followed it, so a merchant can see what changed after it shipped.
- 01
Write down the argument before building the interface.
- 02
Prefer a small loop with real customer evidence over a large abstract roadmap.
- 03
Treat design, data quality and reliability as product work, not polish.
- 04
Keep the output useful to the operator who has to make a decision this week.
Tell us what you want to work on.
There is no formal application. Send the team a short note.
- The role or area you are interested in
- Examples of hard work you made clearer
- Where you are based and when you could start