Roblox AI Fracker

Fracking is our term for repeatedly attacking a problem until you find an idea good enough for us to implement.

You are given a problem.

For example:

• A game's thumbnail CTR is too low

• Players are dropping out at a certain point

• An icon is not communicating clearly enough

• A progression system feels weak

• A monetisation mechanic is underperforming

• A feature technically works but players do not care about it

Your job is not to brainstorm once and give us five suggestions.

You keep going.

You research the problem. You use AI. You study competitors. You look at player behaviour. You challenge the assumptions behind the problem. You generate ideas, discard bad ones and generate more.

You keep fracking until you find something that makes us say:

"Yes. We should actually implement that."

That is a successful frack.

Every successful frack earns you a $500 bonus.

The idea does not need to ultimately succeed after implementation. Games are experiments and sometimes good hypotheses still fail.

What matters is that your thinking was strong enough for us to commit time and resources to implementing the idea.

Over time, the goal is for a Fracker to become increasingly accurate.

Not just better at generating ideas.

Better at identifying which ideas are actually worth doing.

Yes, we are hiring a Roblox AI Fracker.

This is not a Producer role. It is not a traditional Game Designer role. You are not being hired to manage people, organise tickets or spend your day in meetings.

You are being hired to think with AI.

We operate several large Roblox games and there are constantly hundreds of things that could potentially be improved.

CTR. Thumbnails. Icons. Retention. Monetisation. Progression. Features. UX. Game loops. Social mechanics. Discovery. Small interactions. Strange problems nobody has properly investigated yet.

There are effectively unlimited problems.

The limited resource is high-quality human attention.

Your job is to apply that attention continuously.

You will spend your day fracking problems across our games.

You will use AI heavily. You will research. You will generate ideas. You will challenge assumptions. You will learn how our Creative Director thinks and over time develop your own intuition for what will actually work.

Most ideas will go nowhere.

That is expected.

The goal is not to generate the largest number of ideas.

The goal is to generate accurate ideas that are good enough to implement.

When one of your ideas gets implemented, that is a successful frack.

Every successful frack earns a $500 bonus.

• Receive specific problems from our live Roblox games

• Investigate why those problems exist

• Generate potential solutions using AI, research and your own judgement

• Frack problems around CTR, thumbnails, icons, retention, monetisation, progression, UX, gameplay and other areas

• Study competitors and wider Roblox platform behaviour

• Generate significantly more ideas than we could realistically implement

• Ruthlessly eliminate weak ideas

• Keep attacking a problem when the obvious answers are bad

• Present your strongest solutions clearly

• Learn directly from our Creative Director's decision-making and intuition

• Track which ideas are implemented

• Study what happens after implementation

• Improve your judgement based on actual outcomes

• Eventually challenge our assumptions when your judgement becomes strong enough

• Teach future Frackers once you have demonstrated consistently strong performance

Imagine we tell you:

"This game's thumbnail CTR is not good enough."

A normal approach might be to create three new thumbnails.

A Fracker goes much further.

You might analyse successful thumbnails across Roblox. Investigate competing experiences. Examine character size, composition, facial expressions, colour contrast, text usage and visual hierarchy.

You might generate 50 concepts with AI.

Then reject 45 of them.

You might realise the thumbnail is not actually the core problem and that the game's positioning itself is unclear.

You might identify a completely different visual concept nobody internally considered.

You present the strongest option.

If we decide:

"Yes. Put that into production."

That is a successful frack.

You earn $500.

Then you move onto the next problem.

We deliberately keep the definition simple.

A frack is successful when:

You propose an idea and we decide to implement it.

That is the trigger for the bonus.

You do not need to wait months for perfect performance data before getting rewarded.

If we believe your idea is good enough to commit development, art or operational resources to it and we implement it, you have successfully fracked the problem.

Successful frack = $500 bonus.

Implementation does not guarantee that the idea ultimately succeeds.

Games are unpredictable.

An experiment can be sensible and still fail.

Your job is therefore not to magically predict the future.

Your job is to continuously produce ideas with a high enough expected value that we believe they are worth testing.

Over time, we expect your accuracy to improve.

A strong Fracker should eventually reach a point where their judgement carries significant weight because they have repeatedly demonstrated that their ideas are worth implementing.

You must be Roblox native.

That does not mean you need to be an elite Luau programmer.

It means you genuinely understand Roblox.

You understand the platform, players, culture and what makes Roblox games work.

You should instinctively understand concepts such as discovery, CTR, retention, progression, monetisation, social systems and why Roblox players behave differently from traditional PC or console audiences.

You might currently be a:

• Roblox developer

• Game designer

• Product-minded Roblox creator

• LiveOps designer

• Roblox analyst

• Growth-minded developer

• Experienced Roblox creator

• Extremely obsessive Roblox player who has also built things

• Someone difficult to categorise who is unusually good at solving game problems

We care much more about how you think than your current job title.

• You are extremely curious

• You naturally attack problems from multiple directions

• You can spend hours thinking about one seemingly tiny issue

• You use AI as a thinking multiplier rather than a replacement for thinking

• You understand Roblox deeply enough to recognise bad platform advice

• You generate a lot of ideas but are ruthless about discarding weak ones

• You care whether something actually works rather than whether it sounds clever

• You are competitive

• You like measurable output

• You like the idea of earning directly from strong ideas

• You actively want your judgement tested against reality

• You can take direct feedback without becoming attached to your ideas

• You are comfortable being wrong repeatedly

• You become increasingly effective as you accumulate context

• You need a clearly defined checklist every morning

• You want somebody to tell you exactly how to solve each problem

• You believe brainstorming itself is valuable regardless of output

• You produce huge strategy decks instead of answers

• You have never seriously used AI

• You do not understand Roblox

• You are uncomfortable having your ideas rejected constantly

• You care more about owning an idea than finding the best solution

• You prefer maintaining processes over solving problems

• You want a role where performance is difficult to measure

You receive a base salary.

On top of that:

Every successful frack earns you $500.

The base salary exists because experimentation is inconsistent and we do not expect every idea to work.

The bonus exists because we want the incentive to remain extremely clear:

Find things worth doing.

There is no expectation that every problem produces a successful frack.

Some problems will take hours.

Some may take days.

Some will produce nothing useful.

Others may produce multiple ideas that we decide to implement.

The objective is not to optimise for cheap quantity.

It is to continuously improve your accuracy.

As the team grows, Frackers will learn from one another, compare approaches and continuously increase the accuracy of the group.

The long-term objective is for the combined judgement of the Fracking team to eventually exceed the intuition of the people currently making those decisions.

We are starting with Roblox.

But Fracking is ultimately a system for allocating more intelligent attention to problems.

One Fracker learns how we think.

Several Frackers learn from each other.

The strongest Frackers teach new ones.

The team develops institutional knowledge around which approaches work and which do not.

Eventually you have an entire group spending every working day attacking problems that otherwise would never receive enough attention.

The goal is not an enormous brainstorming department.

It is an accuracy department.

A human and AI problem-solving engine whose sole purpose is to continuously identify better decisions.

Think of it as bitcoin mining for hit-making.

Instead of mining hashes, you are mining for ideas worth implementing.

Every time you find one, you earn $500.

We call that Fracking.

Do not send us a generic cover letter explaining that you are passionate about games.

Tell us:

1. Your Roblox experience

2. Which Roblox games you understand particularly well and why

3. The AI tools you currently use and how you use them

4. One thing you believe a successful Roblox game is currently doing badly

5. What you would change and why

If your answer to number four is interesting, we will probably have something to talk about.