### About Parker
Parker is the AI marketing brain for every business in the world.
We launched in February and went viral. Parker now advises on over $1 billion
in annual ad spend across hundreds of brands, including some of the largest
consumer companies in the world.
See our launch here:
<https://x.com/heyparkerdotai/status/2016221027093905561?s=20>
It's never been easier to start a business. But attention is limited. Every
business owner alive is fighting for the same shrinking pool of customer
attention, and most of them are losing. They guess at messaging. They copy
competitors. They burn cash on ads that die in three days. They have no idea
what their customers actually want to hear.
Parker sits in the middle of that fight. We help business owners win attention
efficiently across paid, organic, retention, forecasting — every surface of
the marketing funnel. The marketing function at every company on earth is our
addressable surface area. If we win this position, Parker is a generational
company.
The work spans the deepest parts of the technical stack and the deepest parts
of human psychology, at the same time. On the technical side, we're building
agentic AI systems that ingest millions of signals — customer reviews, ad
performance data, organic content, comments, transcripts — and turn them into
creative that converts. We're doing original research to imbue our models with
real creative taste. Not just pattern-matching on customer language, but
internalizing the principles of great creative strategy. We want Parker to
have judgment, not just recall.
On the psychology side, we're studying what actually moves people. Why a
phrase like "dad bod" sells more jeans than "athletic fit." Why some ads stop
the scroll and others die. The work sits at the intersection of LLM research,
marketing science, and consumer behavior, and almost no one is operating at
this intersection seriously.
The team behind Parker is rare: top engineers, world-class creative
strategists, and seasoned ad agency operators all under one roof. We're backed
by OVO Fund, Unlock VP, and Hustle Fund, alongside operators from OpenAI,
Anthropic, Meta, Google, and Vercel.
* * *
### The vision for this role
This is a player-coach role. You'll spend meaningful time in the code —
shipping features, architecting systems, reviewing PRs — while also leading
the team and setting technical direction. You're hands-on, not hands-off.
Parker today is great. Parker tomorrow is true AI agent orchestration —
autonomous agents that research customer trends, write ad strategy, generate
creative, run tests, and double down on what's working, all in concert. No
human in the loop except to set direction.
Most of the industry is still gluing together prompts and calling it an agent.
We're past that. You'd own the architecture of our agentic systems — the
orchestration layer, the eval harnesses, the memory systems, the failure
modes. The work is hard and there isn't a playbook for it.
There's also a research layer. We're finetuning our own models on creative
taste — what language converts, what stops the scroll, what doesn't. Real ML
research, with a short line to product. You'd lead it.
The bigger bet: we want to keep the engineering team small and have AI agents
do most of the dev work. The team's job is managing those agents, not writing
every line. Engineers as AI managers. Very few companies are actually running
this way. You'd be the one setting us up to.
* * *
### Why this is an engineer's dream
A small AI-leveraged engineering team is only half the story. The other half
is distribution. Half our team is made up of world-class ad agency founders
with serious reach into the market. That means everything you build lands in
the hands of real customers immediately. No sitting on shelfware. No begging
for design partners. You ship it, and it goes straight to brands spending real
money to grow.
Most AI engineering roles give you the tech but not the distribution. Most
adtech roles give you the distribution but not the AI. Parker is both, and the
combination is rare.
* * *
### What you'll own
* Reporting directly to the CEO (a Stanford AI engineer by trade)
* Technical leadership of the engineering team (currently 5 people) — setting direction, running hiring, growing people
* Technical leadership of our non-human AI engineering team — the agents doing increasing amounts of the actual dev work, and the systems that let them
* Architecture of Parker's AI agent systems — orchestration, eval harnesses, prompt infrastructure, memory, and the pipelines that power them
* ETL and data pipelines processing large volumes of customer review, ad performance, and organic social data
* Infrastructure reliability and scale — we're growing fast and the product needs to keep up
* End-to-end product delivery across frontend, backend, and API orchestration — including iterating directly with customers, taking their feedback into the code, and shipping improvements fast
* Using the latest AI coding tooling (Claude Code, Cursor, etc.) to ship at 10x speed — this is how we work, not optional
* * *
### Our stack
TypeScript end-to-end. Mastra SDK for agent orchestration, Supabase, Qdrant,
Redis, Temporal, Langfuse, Vercel AI SDK, GCP. We use whatever the best tool
for the job is — often something released in the last six months.
You don't need to have used all of these, but you should be deeply comfortable
with the patterns: agent orchestration, vector search, durable workflows, eval
and observability for LLM systems, scalable architecture, and the devops side
— CI/CD, infra-as-code, monitoring, incident response.
* * *
### What we're looking for
**AI-native engineer.** You've built production multi-agent systems and
complex LLM orchestration — not just prototypes. You understand prompt
engineering, agent architectures, RAG pipelines, and the tradeoffs between
them.
**Experienced architect.** You've designed and scaled systems that handle real
load. You know how to build ETL pipelines, design for reliability, and make
infrastructure decisions that don't need to be unwound in six months.
**Strong IC who leads.** You write code every day and you're great at it. You
also know how to set technical direction, run a team, and make people around
you better.
**Strong communicator.** You can explain complex technical decisions clearly
to anyone in the room — engineers, customers, investors, the rest of the team.
You're comfortable talking with customers directly and folding what you hear
back into the work.
**AI tooling power user.** Claude Code with sub-agents and custom MCP servers,
Cursor with full repo context, Codex, Devin, Windsurf — whatever the frontier
looks like the week you read this. You're running multi-agent workflows,
building your own internal tooling on top of these systems, and you expect
your team to do the same.
**Startup-minded.** You've either founded something, been an early employee at
a fast-growing startup, or otherwise know what it feels like to build with
urgency and limited resources. You don't need a playbook.
**Plugged into dense talent networks.** You came up somewhere that puts you
near great people — a top school, a well-known tech company, a serious
research lab. Hiring is going to matter, and the people you already know are
part of how we win.
* * *
### Bonus points
* You've led a small eng team through a period of rapid growth
* You have strong opinions about eval infrastructure and test harnesses for AI systems
* You've worked in performance marketing, adtech, or a data-heavy consumer product
* * *
### What this is not
* A pure people-manager job. You'll be in the code.
* A corporate engineering leadership position. We move fast and expect you to.
* * *
### To apply
**Send us two things:**
1. **Introduce yourself.** Who you are, why this role interests you, and why you'd be great at it. Specific beats long.
2. **A short written answer (a few sentences) to one of these:**
* What's the most interesting agentic system you've seen built recently, and what makes it good?
* What do you think Parker should build next?
If you have a personal site, Twitter, or GitHub, include a link. Resume
optional — attach if you want.