AI Developer

Ocean Park Asset; Santa Monica, Northern (CA, KY); Hybrid; USD 262,000 - 441,000; Part time

Benefits offered by this job

Fully remote

Flexible hours

Freelance / contract

Immediate start

Job summary

Ocean Park Asset is seeking an AI Developer to write the decision policy our systems actually run. You will output policies that decide enter/hold/exit for a single position, with no leverage and measured risk.

The role emphasizes live-loop maintenance, plain-language explanations of risk walls, and keeping policy and risk constants intact while improving the system for real capital use. Remote collaboration is core to the engagement.

Qualifications

• You have built control logic, trading rules, or another system where a bad decision has a number attached.
• You treat drawdown, position size, and costs as part of the design.
• Comfortable working remotely, mostly async, with high autonomy.

Responsibilities

• Write decision policies that ignore noise and act when the move is real
• Size a position so a gap cannot break the loss limit
• Exit when the reason to be in is gone, or when risk is threatened
• Keep the live loop and any research behavior from drifting apart
• Explain, in plain language, why the risk wall still holds
• Leave the engine and the risk constants alone. Fix the policy, not the ruler

Skills

Python

Job description

AI Developer

$190 – $320 USD / hour

Depending on experience. Open to a longer collaboration if the first one is good

The decision policy. ENTER, HOLD, or EXIT. Not a language model.

We are looking for an AI Developer to write the decision policy our systems actually run. A price loop is already in motion. On each tick your code returns enter, hold, or exit. One position. No leverage. Risk is measured, not suggested — a limit that is not in the decision is not a limit. Size matters more than a clever exit, because a gap can hit before your next chance to get out. If that is the kind of AI work you want, you will like it here. There is no model to fine-tune, no prompt to chain, and no chatbot to wrap.

What you'll do

• Write decision policies that ignore noise and act when the move is real
• Size a position so a gap cannot break the loss limit
• Exit when the reason to be in is gone, or when risk is threatened
• Keep the live loop and any research behavior from drifting apart
• Explain, in plain language, why the risk wall still holds
• Leave the engine and the risk constants alone. Fix the policy, not the ruler

What we're looking for

• You have built control logic, trading rules, or another system where a bad decision has a number attached
• You treat drawdown, position size, and costs as part of the design
• You will not 'fix' a failing policy by editing the limits
• Comfortable working remotely, mostly async, with high autonomy

Nice to have

• A live or paper strategy you have had to turn off
• You have sized a position for a gap, not only for the average day
• Python research someone else can rerun, if that is how you think
• Enough market intuition to know noise from a move, and enough humility to encode it

Engagement

• Freelance / contract, with potential for long-term collaboration
• Flexible hours
• Fully remote
• Immediate start

Rate range

$190 – $320 USD / hour

Depending on experience. Open to a longer collaboration if the first one is good

• Real ownership. You ship the system you design. No committee between you and the diff.
• The code sits in the path of real capital. Not a demo that gets thrown away after the pitch.
• Fully remote, from day one. There is no head office you are missing out on.
• We pay for one excellent person, not three average ones. The rate on this page is the rate.

Hiring process

Our hiring process involves:

• A two-hour practical inside an existing price loop. You ignore noise, act on a real move, and keep a written risk limit intact — including when price gaps. We do not score profit. We read whether the policy is honest.
• A technical conversation with engineering
• Final conversation and contract for selected candidates