AI Engineer

What Youll Own:

• Design production-ready GenAI pipelines:

  Build and maintain node-graph and script-based workflows using

  ComfyUI, Stable Diffusion, and modern image/video models

  for key use cases such as keyframes, style transfer, image-to-video, text-to-video, upscaling, and repair.

• Ensure visual consistency across shots

  Design

  first-frame-controlled and reference-guided workflows

  that preserve characters, composition, lighting, and style across multiple frames and sequences.

• Bridge creativity and engineering

  Collaborate closely with

  GenAI artists, producers, and creative leads

  to translate creative intent into stable, reusable workflow templates.

• Optimize for speed, cost, and quality

  Profile and optimize pipelines through smart node selection, batching strategies, caching, and model loading optimizations—ensuring maximum GPU efficiency and minimal inference cost.

• Expose workflows as services

  Work with the

  Technical Director

  to package pipelines as

  APIs and internal services

  , with clear contracts, versioning, and documentation for easy integration by other teams.

• Experiment and evolve

Continuously evaluate new models, nodes, and techniques (LoRA, ControlNet, IP-Adapter, video diffusion tools) to improve reliability, quality, and creative flexibility.

What You Bring:

• 1-2+ years of experience in

  technical media, VFX pipelines, creative coding, ML, or computer vision engineering

  .
• Strong hands-on experience with

  ComfyUI or similar node-graph systems

  , including deep familiarity with

  Stable Diffusion ecosystems

  .
• Solid understanding of

  LoRA, ControlNet, IP-Adapter, reference conditioning, and modern image/video generation models

  .
• Strong

  Python

  skills and experience using

  ffmpeg

  for programmatic image and video processing.
• Working knowledge of

  GPU resource management

  , inference optimization, VRAM constraints, and common generative model failure modes.
• Ability to think in systems—turning one-off experiments into

  repeatable, production-grade workflows

  .
• A portfolio of

  workflows, tools, demos, or repositories

  (GitHub or similar) is highly desirable.

Why Obbserv AI:

building the workflow layer of the future

architect the pipelines that turn imagination into repeatable reality

If you’re excited about building the invisible engines behind the most visible creative work this is where you belong.

Welcome to Obbserv AI.