Founding Applied AI Engineer

Department: Product Engineering

Location: San Francisco

Compensation: $200K – $250K • Offers Equity

Employment Type: FullTime

TL;DR: We're building an AI mentor that knows you deeply and pushes you to grow. Not a chatbot. Not an assistant. A mentor that observes your life, understands who you are, and steps in when you need it - whether you ask or not. Just raised $6M pre-seed. Looking for 3 founding engineers to build this with us in SF.

Note: This job description was written by Claude Code, using our founders' WhatsApp conversations as context. We practice what we preach: building with AI.

What we're actually building

An AI that builds a deep psychological profile of you and uses it to push you to grow. Not a chatbot. A mentor. One that watches how you spend your time, understands who you are, and steps in when you need it, whether you asked or not.

The hard technical problems aren't the ones you've seen before:

• Psychological profiling from a digital footprint — how do you actually understand a person from their emails, calendar, messages, and behavior patterns? Not just extract facts. Understand them.

• Memory that compounds — most AI products have terrible memory. Context windows are finite. Vector DB retrieval is mediocre. We believe file-based context engineering inspired by coding agents beats existing approaches for our use case and we're building to prove it.

• Proactive, agentic behavior — a mentor that only responds when you ask isn't a mentor. Building AI that initiates, at the right moment, in the right way, is an unsolved product and engineering problem.

• Morphing UI personalization — the same AI, completely different behavior and interface depending on who you are. Making that work at scale without it feeling inconsistent is genuinely hard.

We're not doing AI research. We're the applied team that takes what's emerging from labs and ships it as product before anyone else.

What you'll own

• The psychological profiling engine — the system that actually understands users from their data. Architecture, models, context engineering, everything

• Memory architecture — going beyond RAG and vector DBs to build context systems that get smarter over time. File-based, structured, compound

• Agentic orchestration — multi-step AI behavior, tool use, proactive interventions, long-running tasks. Making agents that do things, not just answer questions

• Eval frameworks — how do you know the mentor is actually helping versus just sounding wise? You'll build the systems to measure real behavior change

• Model-agnostic infrastructure — we use Claude today, but we're built to benefit from every lab's progress. The abstraction layer that lets us swap and combine models as the landscape evolves

• LLM API integration — streaming, latency optimization, cost management, prompt engineering, structured outputs
  

You'll be writing code day one. You'll also be making architectural decisions that define the company. We move fast. We ship, see what works, iterate. We use AI to code (obviously), but we're not vibe-coding everything - we own the architecture and fully understand critical systems.

You're a fit if...

Technically:

• You've built real agentic systems — not chatbots, not RAG pipelines; systems where the AI takes sequences of actions, uses tools, manages state across time, and handles failures gracefully

• You have opinions on memory architecture — you know why naive RAG fails for deep personalization and have a point of view on how to do it better

• You're model-agnostic but opinionated — you've worked with Claude, GPT-4, Gemini, open-source models and know when to use which

• You read papers and ship products — you keep up with what's coming out of labs because you want to ship it first, not because you want to do research

• You're frustrated by thin wrappers — a system prompt and a streaming API call isn't going to be enough for what Quest is trying to do
  

Personality:

• Ambitious — You want to build a generational consumer company, not just a feature or a nice product. You're here because you believe Quest can be massive and you want to be part of making that happen

• Creative — You think of things AI won't. You see connections and possibilities that models don't surface. In the age of AI, creativity is what separates great engineers from prompt operators

• Opinionated — You have strong views on what good code looks like, what good product feels like. We want your proactive recommendations, not just execution

• Autonomous — You identify problems and fix them. You don't wait for instructions. After a first conversation, you're the kind of person who comes back with a prototype or a sharp question

• Fast learner — You can absorb a huge amount of information quickly and think creatively around it. This space moves fast; you move faster

• Product-minded — You understand why technical decisions matter to users. Every architecture choice flows from a product idea

Preferred:

• Python expertise, you're fast and your code is maintainable

• Strong CS fundamentals, you understand what's happening underneath the frameworks

• Experience with Claude's API, tool use, extended thinking, or the Claude Agent SDK

• Background in behavioral science, psychology, or persuasive technology — not required, but you'd bring it immediately to how you think about our core problem

• Side projects that use AI in interesting ways, we want to see what you build when no one's asking you to

Seniority:

We don't care about years or titles. If you're 3 years in and you've built something genuinely impressive with AI, we want you. If you've led AI engineering at a Series B, we want you. Show us what you've built.

What you want:

• Build something people actually use and love (not just abstract acclaim)

• Work with two repeat founders who shipped some of Google's first AI agents and sold their first company to Google

• Be in SF with the team - 4 days/week in-person (we need same timezone, same room for rapid iteration)

• High equity + competitive salary - we believe in a small team that scales

• Small, excellent team. Zero compromise on who we hire. We'd rather take 6 months to find the right person

You're NOT a fit if...

• You want to do AI research, not applied engineering — we respect research, this isn't that role

• You think RAG is sufficient for deep personalization over months

• You haven't shipped AI systems that real users depend on

• You want clearly scoped AI tasks — "improve the prompt," "add a tool call." We need architectural ownership

• You're not excited about the actual product — an AI that pushes people to grow. You need to believe in this

• You want to specialize in AI infrastructure without caring how users experience it

• You're remote-only (we need you in SF 4 days/week)

Founders

Vincent (CEO) - Led AI agent products at Google Labs. Shipped some of Google's first agents: Colab data science agent, Jules (async coding agent), Stitch (AI-first Figma competitor).

Thomas (CTO) - Led senior Android engineering teams at Google. Refactored Android connectivity stack, prototyped always-on AI wearables. Shipped Tempow tech to 10M+ devices.

We are repeat founders. We co-founded Tempow in 2016, sold to Google in 2021, moved Paris → SF. Worked together 9 years.

How we actually work

From Vincent & Thomas's actual working style:

We ship and iterate constantly. "Vibe eval" is real. We move fast

We're casual but serious - informal communication, sophisticated thinking

We own our architecture deeply.

We use AI to code but understand every line that's critical

We make decisions fast once we have conviction - no analysis paralysis

We want to build the best AI applied team

What excites us:

Seeing people actually use what we build and achieve their goals

Solving hard technical problems (memory, agent orchestration, morphing UI)

Building consumer products with taste

Shipping breakthroughs before anyone else

What frustrates us:

Overengineering without clear product need

Window shoppers who aren't committed

People without clear priorities or goals

Bureaucracy and process for process's sake (we left Google partly for this)

Our operating principles

• Applied Pioneers - First to adopt emerging AI research and ship it as product

• Magic-First - Consumer products are only worth building if they feel magical

• Ship and Try - Iterate by shipping code and playing with prototypes, not endless docs

• Small Excellent Team - Zero compromise on hiring. Strong ICs who contribute everywhere

Work conditions

Location: In-person in San Francisco, 4 days a week at the office
Immigrant-friendly: Work remotely from Europe, Asia, Australia, or anywhere for up to a month per year. We gladly sponsor visas and green cards.
Parent-friendly: 3 months fully paid parental leave. Both founders disconnect daily from 5-7pm to be with their kids.

This isn't for everyone. We're building for people who want to grow, and we need people who want to build something that actually matters. If that's you, let's talk.