Video Prompt Engineer Jobs (NOW HIRING)

Loop AI is an agentic restaurant intelligence software that augments the back office of restaurant chains by automating workflows and delivering intelligence across the finance, operations and marketing functions. Loop deploys AI agents built by our in-house team of AI engineers, strategists and subject matter experts into restaurant brands - bringing industry best practices in handling complex internal functions. We have offices in San Francisco, New York, Tampa and India.

Loop is one of the fastest growing restaurant technology companies powering a few billion dollars in revenue and growing to serve 10K+ restaurants within 3 years across some of the most recognizable brands of the USA (McDonald’s, Burger King, Sweetgreen, Dave’s Hot Chicken - to name a few), helping them grow their topline & bottomline.

Loop is built by a world class team of entrepreneurs, operators, leaders and AI engineers from different industries, ranging from cutting edge big-tech, management consulting, investment banking among others across companies like Uber, Google, Amazon, McKinsey and others.

You're here to define how Loop talks to its customers on screen - and to build the AI-leveraged engine that does it at the pace we ship product.

Loop's products power a few billion dollars in delivery revenue and are landing in the hands of tens of thousands of restaurant operators. Most of them don't have the time to read through docs - they learn through video, on a phone, between shifts, with a delivery tablet in the other hand. The version of customer education that wins for Loop is video-first. The version of that system that works at our scale is templated and AI-leveraged.

The person we want thinks in three lenses at once:

• The base - the design language and clip library that everything is built off

• The pipeline - the workflow (AI-leveraged or otherwise) that turns idea into finished video in hours/days, not weeks

• Scalability - the system that survives weekly product change and scales to hundreds of customers without rebuilding from zero

You'll be working with the agent owners, the GMs, and the CSMs - figuring out what an operator actually needs to see, then templatizing and pipelining it so we never re-invent.

• The base. Loop's video design language - visual style, brand voice on camera, sound, reusable clip library, templates for the recurring video shapes (onboarding, training, per-customer value, agent walkthroughs).

• The pipeline. The AI-leveraged production stack - pair the templates with the right model and editing layer (After Effects, Remotion, Runway, Veo, Kling, Claude skills, Clueso, etc.) so a per-customer video takes hours/days, not weeks.

• The translation. Take raw context - synopses, pod owner briefs, customer call recordings — and turn it into the script, storyboard, and finished video. The leap from "what we want to teach" to "what an operator actually finishes watching" is where this role exists.

• Scalability. A workflow others can run - standards docs, prompt libraries, model handoffs, the rebuild-resistant approach that survives Loop shipping product weekly.

• The proving set. 15-20 on-brand 60-90 sec explainers in the first phase (30 days) 

• Produced product or brand video at a SaaS / consumer / agency context where the bar was high and the cadence was real 

• Built a template / system / design language for video - not just shipped individual assets. You've already proved you can scale yourself

• Fluent across the traditional (After Effects, Premiere Pro) and modern AI-video stack - Remotion, Runway, Veo, Kling, Claude skills, etc.. You have a point of view on when to use which (programmatic vs. gen video vs. voice clone vs. editing assistant) and you've shipped on it. Sometimes the answer is After Effects. Sometimes it is Remotion. Sometimes it is a clean screen recording with smart annotations. Sometimes it is a reusable template that can generate 50 customer-specific variants. You should know the difference.

• Worked in-person with product and GTM teams to extract context and translate it into video - fast

• Held the quality bar against deadline pressure. The answer to "ship it ugly" is a no from you

• Contract: 2 months to start, extendable based on output. Open to a longer term if it works on both sides.

• Hybrid. The work depends on being in the room with the people (at least a few in-person sprints) who own the product surface and the customer calls. 

• Compensation. Tell us what you need to do it well - rate, tools, model credits - and we work back from the outcome.