LaunchGood · Full Time · Remote

We're hiring our first Applied AI Engineer. Instead of a resume, build us something.

LaunchGood is the global crowdfunding platform for the Muslim Ummah, used by people across 130+ countries to fund thousands of campaigns every year. Behind every campaign that gets funded, a small, values-driven team is doing the unglamorous work of moderating, supporting, reconciling, and analyzing. That work scales linearly with our growth. AI doesn't have to.

We're putting an Applied AI Engineer on the team to change that math. We want to evaluate you the way the job will actually be done.

Your CV isn't everything

We're looking for people who look at a broken process and can't help but imagine what it should be. You will need to move fast in ambiguity, own problems end-to-end, and think carefully about where AI should take responsibility, and where it shouldn't.

Your story tells us where you've been. Your build tells us what you can do.

The challenge: Prototype a System

Design and prototype an AI system that meaningfully expands what a small, mission-driven team can do.

Pick a real problem. The kind a 10-person ops, support, finance, or trust & safety team might actually face. Then show us a working system that takes repetitive work off humans, surfaces insight they didn't have, or helps them serve more people without losing quality.

Need Ideas?

• Systems that gather data from multiple sources to synthesize information for human review
• Upstream reporting capabilities, e.g. gather data from multiple sources to present as dashboards or regular reports
• Workflows which require human-in-loop before enabling actions which modify core systems (e.g. sending an email)
• Moving data from one system to another, or mapping it between systems

You don't need to know how LaunchGood works internally. Pick a problem you understand. Make reasonable assumptions and state them explicitly. We care about how you think, not whether you guessed our roadmap. Problems that are at the intersection of multiple teams tend to be much more complex and where we want to see how you make your trade-offs.

What to submit:

• A working prototype. Link to a deployed demo. It has to actually run. Do not submit actual code. We want to only see deployed demos.
• A short video walkthrough (≤5 minutes) showing the system in action and explaining your key decisions, we specifically want to understand how human/AI interact, and what are some of the trade-offs you had to focus on. 
• You can submit the above as links in a PDF document or include them under ‘Additional Information’

What we're evaluating:

• Problem choice. Is this a real problem worth solving, or AI for AI's sake?
• Human/AI boundary. Have you thought carefully about what the AI owns, what stays with the human, and how trust is built between them?
• AI responsibility. Does the AI actually do meaningful work, or is it window dressing on a deterministic system?
• Judgment under uncertainty. How do you handle failure modes, edge cases, evaluation, and cost?
• Systems thinking. Would this hold up in the real world at scale, or fall over the first time data gets weird?
• Communication. Can you explain your reasoning clearly to someone who has to bet on you?

You should use AI to build this. We expect you to. What we care about is the system you design, not how much help you used getting there.

What the role actually is

You'll report directly to the Head of Engineering and partner with team leads across LaunchGood to find and ship the highest-leverage AI opportunities. The mandate is broad. Take the repetitive, judgment-light, or insight-starved parts of how LaunchGood runs internally across ops, trust & safety, support, finance, marketing, and engineering. Rebuild them with agents, copilots, and automations that work in production.

This is not a research role. You'll ship things people use every week, measure whether they save real time and improve real quality, and iterate fast. You'll have wide latitude to pick the right tool: a Python agent with custom tools and evals, a TypeScript service, or a well-designed n8n or Zapier workflow with an LLM step in the middle. Pragmatism wins.

In your first 3 months, you'll have:

• Mapped at least 5 highest-leverage internal workflows across LaunchGood with team leads and built a prioritized roadmap.
• Support internal tools and systems that enable AI adoption across LaunchGood
• Shipped production agents or automations that have measurably saved hours or improved quality for at least two internal teams.
• Stood up an internal copilot over LaunchGood's institutional knowledge that staff actually use daily.
• Established an internal eval framework so any AI work (yours or others') has a clear, repeatable bar for "is this working?"
• Led workshops for internal enablement programs.

What we'd expect you to be comfortable with on day one:

• Python and/or TypeScript, comfortable across backend and lightweight internal UIs.
• Modern LLM tooling: model APIs (Anthropic, OpenAI, others), RAG, vector stores, structured outputs, tool use.
• One or more agentic frameworks (LangGraph, CrewAI, Pydantic AI, AutoGen, or equivalent), with an opinion on when not to use them.
• Designing evals (deterministic, LLM-as-judge, and human review) and wiring them into CI.
• Low-code workflow platforms (n8n, Zapier, Make), and knowing when to reach for them vs. writing code.
• Integrating with the messy real world: Slack, Google Workspace, internal databases, CRMs, payment systems.
• Sound judgment around data privacy, sensitive content, and community trust.

FAQs

• What is this exactly? A full-time hire for an Applied AI Engineer role at LaunchGood, with an unconventional application: instead of a resume, you submit a working AI prototype.
• Who should apply? Builders who think in systems, move quickly under ambiguity, and care deeply about where AI should and shouldn't replace human judgment. You take ownership end-to-end. We care more about proof of building than titles or tenure.
• Is this an internship or part-time? No. Full-time role, reporting to the Head of Engineering.
• Where is the role based? Fully remote. LaunchGood is a globally distributed team.
• Do I need to live in a specific country to apply? No, but you need to be legally allowed to work in your current location.
• Do I need to know how LaunchGood works internally? No. We're evaluating how you think: how you frame problems, what assumptions you bring, and how you design under real-world constraints. If something is unclear, make a reasonable assumption and state it explicitly.
• How technical does the application need to be? Technical enough to demonstrate real capability and reasoning. If the AI never actually does anything, it's probably not enough. It cannot be a prototype alone, it needs to leverage real data (can be mocked). 
• Can I use AI tools to help build this? Yes. We expect you to. What matters is the system you design, not how much help you use.
• How do you select? We look for sound judgment and systems that would hold up in the real world: a real problem, a clearly defined human/AI boundary, meaningful AI responsibility, and clear communication. We pay attention to how you handle uncertainty, pressure-test failure modes, design for scale, and explain your reasoning.
• When do I find out? Applications are rolling. Submissions will be typically reviewed within 48 hours. Strong candidates will be contacted for panel interviews immediately as well as our recruiting team.
• What happens to my submission? Your submission will be used solely to evaluate your application. All application materials are stored securely and permanently deleted 30 days after the close of the hiring process.

About working at LaunchGood

We're a fully remote, globally distributed team united by a simple belief: Muslims are incredible, and the world is better when their ideas get the support they deserve. We offer flexible hours, Eid bonuses, team retreats every year (check out our team’s highlights in Bosnia here), and the unusual privilege of doing work where values and career meet.

We'd love to see what you build