Gen AI Model Artist

2 hours ago

London, United KingdomSenior

Responsibilities

Train and fine-tune generative AI models for hyperrealistic synthetic image and video media.
Curate, balance, filter, caption, de-duplicate, and prepare datasets for generative training.
Iterate on fine-tuning strategies, including adapter and LoRA-style approaches, to improve realism and likeness.
Diagnose identity drift, artifacts, texture issues, uncanny results, overfitting, and prompt or conditioning sensitivity.
Manage training runs, evaluation sets, regression checks, comparisons, and experiment logs.
Evaluate facial likeness, realism, consistency, and production readiness.
Collaborate with R&D and cross-functional teams to develop and integrate improved GenAI techniques.
Monitor and optimize inference speed, VRAM usage, throughput, consistency, and overall model quality.
Prototype emerging GenAI tools and workflows and translate creative or production needs into technical solutions.

Proven experience training and fine-tuning generative models, including Stable Diffusion-family or similar diffusion, GAN, or transformer-based models.
Strong GenAI dataset skills, including curation, balancing, subject coverage, filtering, and evaluation-set preparation.
Familiarity with UNIX/Linux environments and GPU training workflows.
A critical eye for facial likeness and visual quality, with the ability to explain issues clearly.
Strong teamwork, communication, problem-solving, and creative-thinking skills.
Ability to remain calm under pressure and meet deadlines.
VFX or compositing experience, especially facial or beauty integration, is preferred.
Python and ML engineering experience in training automation, data pipelines, or evaluation tooling is preferred.
Experience with face-swap systems and face workflows is preferred.
Experience with Nuke or similar compositing tools is preferred.

5,001-10,000 employees