Introducing ONOMA AI
At ONOMA AI, we harness technology to empower creators through advanced generative AI solutions. Our proprietary text-based generation pipelines craft imaginative stories, while our in-house image generation models bring these narratives to life with stunning visuals. We are dedicated to enhancing storytelling and visual expression, providing seamless AI-assisted tools that support creators across various industries. Join us in shaping the future of creativity with cutting-edge innovation.
About the job
As a Part-Time AI Research Team Member, you will contribute to cutting-edge advancements in generative AI, specializing in text-to-image models, multi-instance consistency, adapters, and inference methods. Your role will be tailored to your interests and expertise, focusing on one or more of the following areas: researching innovative AI techniques, developing new methodologies, and collaborating with like-minded experts to push the boundaries of AI-assisted creativity. Whether you are actively engaged in the image generation community, an experienced researcher, or an academic with expertise in the field, we welcome your unique skills and insights.
What You’ll Do
Research on Text-to-Image Models
Conduct research on text-to-image models such as Diffusion, CFG, LDM, DIT, and AR.
Analyze the limitations of current image captioning and augmentation methods.
Explore advancements in samplers, schedulers, VAEs, and related techniques.
Research on Multi-Instance-Consistent Generation
Research on Adapters
Research on Inference Methods
Optimize upscalers, ControlNet, Adetailer, RefDrop, and similar techniques.
Conduct research on image preprocessing, postprocessing, and enhancement strategies.
Explore optimization and lightweighting techniques, including cross-attention optimization, Hypertile, and Forge.
Develop applied methodologies such as line art extraction, coloring, and layer separation.
We are looking for people who
Actively participate and contribute to the image generation community and fundamental programming skills necessary for conducting research
Passionate about and/or interested in manga and can keenly observe subtle differences in story and artwork
Stay true to the fundamentals while enjoying discussions and problem-solving for new challenges
Listen attentively to colleagues and engage in conversations with sincerity and respect
Approach problem-solving proactively and is skilled at finding solutions through diverse methods
Eager to learn new technologies and take joy in sharing knowledge with others
Possess perseverance and strong execution skills to see tasks through to completion
Preferred Technical Qualifications
Experience in image generation and adapter training using WebUI, ComfyUI, Kohya ss, etc.
Active participation in model hubs and image generation communities
Hands-on experience in training and deploying image generation AI in real-world applications
Published papers in major international conferences related to image generation AI (CVPR, ICCV, ECCV, NeurIPS, etc.)
Strong analytical skills in animation composition and visual storytelling
How to Apply
A resume
A CV outlining your experience with each AI model hub platform, including your account name (if you’re comfortable sharing)
Please send your application to contact@onomaai.com
Job Location
Recruitment Process
Document Screening
Preliminary Evaluation (Phone interview, coding test if necessary)
First Interview
Second Interview (Includes Culture Fit assessment)
Compensation Discussion / Pre-employment Health Check
Final Hiring & Onboarding
The hiring process may be subject to change depending on the schedule and circumstances.
Additional Notes
The coding test conducted during the preliminary evaluation is used as a reference in the first interview but does not determine pass/fail.
If any false information is found in the application, the offer may be revoked.
If a non-compete lawsuit is filed during the probation period, the employment offer may be canceled.
Interview schedules will be individually communicated to candidates who pass the document screening.
Introducing ONOMA AI
At ONOMA AI, we harness technology to empower creators through advanced generative AI solutions. Our proprietary text-based generation pipelines craft imaginative stories, while our in-house image generation models bring these narratives to life with stunning visuals. We are dedicated to enhancing storytelling and visual expression, providing seamless AI-assisted tools that support creators across various industries. Join us in shaping the future of creativity with cutting-edge innovation.
About the job
As a Part-Time AI Research Team Member, you will contribute to cutting-edge advancements in generative AI, specializing in text-to-image models, multi-instance consistency, adapters, and inference methods. Your role will be tailored to your interests and expertise, focusing on one or more of the following areas: researching innovative AI techniques, developing new methodologies, and collaborating with like-minded experts to push the boundaries of AI-assisted creativity. Whether you are actively engaged in the image generation community, an experienced researcher, or an academic with expertise in the field, we welcome your unique skills and insights.
What You’ll Do
Research on Text-to-Image Models
Conduct research on text-to-image models such as Diffusion, CFG, LDM, DIT, and AR.
Analyze the limitations of current image captioning and augmentation methods.
Explore advancements in samplers, schedulers, VAEs, and related techniques.
Research on Multi-Instance-Consistent Generation
Develop methods to represent multiple objects within a single image.
Investigate approaches for maintaining object consistency across multiple images.
Research on Adapters
Study and enhance adapter-based methods such as LoRA, LyCORIS, and DoRA.
Research on Inference Methods
Optimize upscalers, ControlNet, Adetailer, RefDrop, and similar techniques.
Conduct research on image preprocessing, postprocessing, and enhancement strategies.
Explore optimization and lightweighting techniques, including cross-attention optimization, Hypertile, and Forge.
Develop applied methodologies such as line art extraction, coloring, and layer separation.
We are looking for people who
Actively participate and contribute to the image generation community and fundamental programming skills necessary for conducting research
OR 3+ years of research/development experience in the field
OR Master’s or Ph.D. degree in a relevant field
Passionate about and/or interested in manga and can keenly observe subtle differences in story and artwork
Stay true to the fundamentals while enjoying discussions and problem-solving for new challenges
Listen attentively to colleagues and engage in conversations with sincerity and respect
Approach problem-solving proactively and is skilled at finding solutions through diverse methods
Eager to learn new technologies and take joy in sharing knowledge with others
Possess perseverance and strong execution skills to see tasks through to completion
Preferred Technical Qualifications
Experience in image generation and adapter training using WebUI, ComfyUI, Kohya ss, etc.
Active participation in model hubs and image generation communities
Hands-on experience in training and deploying image generation AI in real-world applications
Published papers in major international conferences related to image generation AI (CVPR, ICCV, ECCV, NeurIPS, etc.)
Strong analytical skills in animation composition and visual storytelling
How to Apply
A resume
A CV outlining your experience with each AI model hub platform, including your account name (if you’re comfortable sharing)
Please send your application to contact@onomaai.com
Job Location
Remote (Part-Time) – Our company is headquartered in Seoul, South Korea.
Recruitment Process
Document Screening
Preliminary Evaluation (Phone interview, coding test if necessary)
First Interview
Second Interview (Includes Culture Fit assessment)
Compensation Discussion / Pre-employment Health Check
Final Hiring & Onboarding
The hiring process may be subject to change depending on the schedule and circumstances.
Additional Notes
The coding test conducted during the preliminary evaluation is used as a reference in the first interview but does not determine pass/fail.
If any false information is found in the application, the offer may be revoked.
If a non-compete lawsuit is filed during the probation period, the employment offer may be canceled.
Interview schedules will be individually communicated to candidates who pass the document screening.