AI Operations Internship, Ramp

• We’re looking for someone who’s not just interested in AI, but obsessed with building workflows — chains of tools + logic + automation that solve gnarly operational problems. Not “ask ChatGPT stuff,” but build systems that allow Ramp to run smoother, faster, with less friction
• Currently pursuing a B.S. (or higher) in a technical or quantitative field OR you’ve done equivalent non-school work (projects, bootcamps, etc.), graduating between 2026 – 2028
• Hands-on experience with data + some tooling: SQL, spreadsheets, APIs, or workflow builders
• Some exposure to prompt engineering or working with AI (even if simple) is a plus — but not enough to just know prompts; we care about how you embed that into workflows
• Clear communicator: you can diagram your workflow, post screenshots, write clear docs
• Empathy for operations: you’ll often be cleaning up the unfinished stuff, making tools people already use less painful

Desirable

• We’re looking for an AI Operations Intern to join us in Summer 2026 in NYC
• As Ramp’s AI Operations Intern, you’ll spend your summer designing, building, and shipping real workflows (often using no-code / low-code tools) that connect multiple pieces — data sources, triggers, AI models, integrations, alerts, etc
• You’ll help us turn messy, manual protocols into reliable, automated flows
• You’ll work directly with cross-functional teams (Product, Data, Engineering), using tools like Gumloop (or similar) to prototype, test, and scale these flows
• Your job is not just to ideate — but to execute
• Architect and execute multi-step workflows: e.g. “When X happens → automatically do A, then B, filter with AI, alert via Slack/Email, log results → trigger follow-up if needed.”
• Use no-code/low-code platforms (Gumloop, Zapier, Retool, etc.) to build flows that combine AI + integrations + data
• Measure workflow performance, find failure points, and improve reliability
• Triaging operational inefficiencies and manual pain points — then building workflows that eliminate them
• Document your flows (so others can trace, debug, reuse) and teach teammates how to use / maintain them
• Experiment: build proof-of-concept workflows, get feedback, iterate, then aim for reliability at scale
• You’ll own real workflows that Ramp depends on — not just toy projects
• You’ll build things that reduce toil for whole teams
• You’ll get to experiment and be creative with which tools to use; we value execution over pedigree