AI Engineer, Agentic Ad Creative (Multimodal)
Core
Build an autonomous AI agent that generates thousands of images and videos daily for Meta, Google, TikTok, Pinterest, Reddit, and Snap, optimizing for CPI, ROAS, and cost-per-asset.
Role type
Senior IC multimodal AI engineer (agentic creative pipeline)
Builds
End-to-end agentic creative pipeline producing platform-ready vertical ads (script, scene plan, b-roll, voiceover, captions) and event-triggered video variants.
Domain
Generative AI (diffusion/video models) + Performance Marketing
Deliverable
production ML models
Required skills
Diffusion model training (LoRA, DreamBooth), Video generation model expertise, Model routing strategy, VLM for policy compliance, Serverless inference deployment, Cost optimization, ComfyUI workflow design, Inference optimization
Preferred skills
VLM pretraining, Open-sourced ComfyUI nodes, Large-scale serverless inference, End-to-end video product shipping
Technologies
Flux, SDXL, Veo, Sora, Runway, Kling, Hailuo, ElevenLabs, Hedra, GPT-4o, Gemini, Claude, Meta, Google, TikTok, fal, Replicate, Modal
Responsibilities
Extend LLM critic loops to include video and copy; Build video pipelines from brief to final ad; Implement generative model router; Develop VLM policy pre-flight checker; Train and deploy brand-style LoRA/checkpoints; Build event-triggered ad pipelines.
Seniority
Senior, hands-on IC
