Founding AI Engineer

Pearl; San Francisco (CA); Hybrid; USD 180,000 - 290,000; Full time

Benefits offered by this job

Real ownership

Hands-on projects

Small founding team

Salary and equity negotiable

Job summary

Pearl in San Francisco is building the Ferrari of travel, delivering personalized experiences through taste-curation models and a production-grade AI agent that plans itineraries and handles bookings for top travelers.

You will own end-to-end systems, improve search and ranking, embed data, and ensure latency, cost and uptime are managed as we scale the platform with a small founding team. Hybrid work and opportunity to shape the product await.

Qualifications

• Experience shipping ML models to production in a real product.
• Strong focus on user experience in luxury travel.
• Ability to own end-to-end system from idea to production.
• Familiarity with retrieval, embeddings, and ranking.
• Experience with recommender systems and geospatial data is a plus.

Responsibilities

• Own the taste and curation models members actually use
• Improve search, maps, ranking, and retrieval so people find the right place
• Embed member, venue, and social data so the system actually understands taste
• Power the AI agent that plans itineraries and completes bookings
• Put models in production. Evals, latency, cost, and uptime are yours. If it is wrong, you see it
• Sit with the product. Talk to members and venues. Cut scope. Ship.
• Be part of hiring and interviewing the next engineers, and help set how this team works

Skills

Taste models

Ranking models

Retrieval systems

Embeddings

Production deployment

Python

TypeScript

Tools

TensorFlow

PyTorch

Docker

Job description

Pearl is building the future of travel, where you can be treated like a regular wherever you go and venues can give personalized service to anyone who walks through the door.

We are the platform reinventing how the world's most discerning travelers discover, share and book places, and how venues tailor service to them.

This is not a mass market product. We are building the Ferrari of travel. We are pre-seed, venture backed, and have more than $2M in credit grants from Google, Anthropic, OpenAI, AWS, and Microsoft.

Pearl learns each member's taste and uses it to curate exceptional places. An AI agent turns a prompt into a considered itinerary, plans the details, and handles the bookings. Members connect with people of similar taste, and that community is itself another source of discovery. Every choice deepens a taste profile that travels with them. Venues get the context to anticipate preferences and cater service. The real-world experience is more personal, and more carefully curated.

Your job is to own the systems that make Pearl feel personal: the taste models that curate places, ranking and retrieval, the taste profile that travels with each member, and the models behind the planning and booking agent. There is no platform team, no PM writing tickets, and no one else to own this in production.

The job

You ship taste, ranking, retrieval, search and maps, and member/venue embeddings end to end. You keep the member's taste profile current as they travel. You put the planning and booking agent's models in production and keep them there. You help set the bar for how we build this, and you hire the next people on it.

What you will do

• Own the taste and curation models members actually use
• Improve search, maps, ranking, and retrieval so people find the right place
• Embed member, venue, and social data so the system actually understands taste
• Power the AI agent that plans itineraries and completes bookings
• Put models in production. Evals, latency, cost, and uptime are yours. If it is wrong, you see it
• Sit with the product. Talk to members and venues. Cut scope. Ship.
• Be part of hiring and interviewing the next engineers, and help set how this team works

What we are looking for

• You are passionate about travel and/or luxury. This is a product that hinges on understanding travelers and hospitality.
• You are an independent builder. You can take a goal or problem, scope it, build it, and put it in production
• You have shipped 0 to 1 models in a real product. Training or fine-tuning, retrieval or ranking, embeddings, evals, and inference in production. Python. Not prompt-only.
• Ideally, you have experience at a startup or working with fast-paced and demanding teams.
• You are AI-forward. You use models and agents as tools inside the product, not as the product
• You care about taste and UX. This is a luxury/premium product, not mass market.
• Remote-first is negotiable, but as we scale we expect to be hybrid and in person in San Francisco
• Recsys, geospatial/maps, agents in production, travel or marketplace data, TypeScript enough to ship with the app team

What this is not

• A research lab or a paper-first job
• A prompt-wrapper role, or a narrow specialist who never ships
• A large, bloated company with lots of red tape

What we offer

• Real ownership and responsibility. You lead projects and own decisions.
• A real seat: models, product, and hiring, not a silo
• A small founding team. No layers
• Salary and equity to be negotiated

Pearl is based in San Francisco.