3-6 Months (with potential for full-time conversion)
Stipend:
Based on skill level and performance
About the Role:
We are looking for a
hands-on AI Intern
with strong proficiency in
diffusion models
and tools like
ComfyUI
to help us build and optimize real-world AI workflows. You will work closely with our product and design teams to implement high-quality image generation pipelines that are consistent, controllable, and production-ready.
Key Responsibilities:
Create and manage
custom ComfyUI workflows
for varied image generation use cases (real estate renders, fashion try-ons, etc.).
Fine-tune and optimize
diffusion model parameters
to achieve desired outputs with architectural and material consistency.
Integrate
ControlNet, LoRA, IPAdapter, upscalers
and other tools to enhance realism and user control.
Implement logic for
batch rendering
,
prompt variation
, and
parameter automation
inside ComfyUI.
Support real-time testing, bug fixing, and performance tuning of pipelines for client delivery.
Requirements:
Solid practical experience with
Stable Diffusion
or other latent diffusion models.
Good working knowledge of
ComfyUI
(or A1111) - including creating and modifying workflows.
Understanding of basic ML concepts: prompt engineering, conditioning inputs, attention control.
Familiarity with ControlNet, LoRA, IPAdapters, and high-res generation methods like tiled diffusion or upscalers.
Comfortable working with visual tools, JSON nodes, and debugging UI-based pipelines.
Good to Have:
Experience deploying image gen pipelines on cloud (AWS/GCP) or local GPUs.
Familiarity with prompt optimization techniques and AI-driven rendering use cases (fashion, architecture, product).
Exposure to Web UIs or automation scripts around ComfyUI.
What You'll Gain:
Real-world exposure to deploying AI image workflows for commercial use.
Opportunity to work on innovative visual AI products used by international clients.
Fast-paced learning with mentorship from a team experienced in AI + creative tech.
Possibility to grow into a full-time product or tech role.