Multimodal AI Engineer

We’re looking for an engineer to design, build, and improve the core AI workflows behind our product.

This is not a traditional ML role. The work is centered on LLMs, VLMs, and image/video/audio generation models used as part of real production workflows. We’re looking for someone who has strong hands-on experience building systems around these models, evaluating them in practice, and improving quality, reliability, and efficiency over time.

What you’ll do

• Design and build end-to-end workflows powered by LLMs, VLMs, and multimodal generation models
• Integrate, manage, and benchmark models for text, image, audio, and video
• Run experiments on prompts, system prompts, model configurations, and inference pipelines
• Build evaluation frameworks using human review, automated benchmarks, and LLM-as-judge style approaches where appropriate
• Analyze model behavior and failure modes, and turn findings into better prompts, better routing, and better workflows
• Develop scoring, ranking, and recommendation layers for multimodal outputs
• Build APIs and internal tools that make these systems reusable, reproducible, and efficient

What we’re looking for

• Proven hands-on experience building applications or internal systems where LLMs, VLMs, or generative media models were central
• Strong understanding of how these models behave in practice, including prompting, evaluation, reliability, and cost/latency tradeoffs
• Experience working with image, video, and/or audio generation models, including evaluating output quality and deciding what works in production
• Strong Python skills and solid software engineering fundamentals
• Ability to design experiments and iterate quickly based on evidence

Nice to have

• Experience fine-tuning or training LLMs/VLMs
• Experience with multimodal retrieval, ranking, or orchestration systems
• Experience building human-in-the-loop workflows for creative tools

About us

We are a startup at the intersection of AI and media. We are building the next generation of tools that help creatives transform stories into production-ready video and reach their fans. Our mission is to be at the service of good stories.