Job Details

Senior ML Engineer with Generative AI experience for an interactive AI avatar system. Tasks include optimizing pipelines, training models, and ensuring character consistency. Remote/hybrid, full-time or project-based.

We are creating a system of interactive AI avatars that communicate with users and can generate images in real-time. The goal is to create stable, "living" characters with a unique visual style, personality, and emotions. We are looking for a strong ML/CV engineer to help take the visual aspect to the next level. Tasks: • Setting up and optimizing ComfyUI / Stable Diffusion / ControlNet / SDXL/Flux/SD3 pipelines; • Training custom checkpoints and refining existing models for product needs; • Tuning LoRA, Textual Inversion, DreamBooth, working with custom datasets; • Creating consistent avatars based on descriptions or references (characters maintain appearance and style between scenes); • Experimenting with various image generation models (SDXL, Flux, SD3, etc.); • Developing and supporting fine-tuning / inference pipelines (VRAM optimization, speed, quality); • Interacting with product and engineering teams to integrate ML into the product (API, UI, etc.).

• Deep experience with ComfyUI or Automatic1111 at a production level; • Understanding of diffusion models and fine-tuning principles; • Experience training custom checkpoints (SDXL / SD 1.5 / Flux / SD3 or similar models); • Experience training LoRA, DreamBooth, Textual Inversion from scratch; • Ability to achieve character consistency (faces, style, poses, emotions); • Experience with ControlNet, IP-Adapter, StyleAdapter; • Strong Python, PyTorch knowledge; • Experience optimizing inference (TensorRT, xFormers, SDXL Turbo, etc.) is a plus; • Bonus: experience integrating ML into production products. Candidate Level: - Strong Senior. Relevant education is a plus.

• Remote / hybrid; • Full-time or project-based work; • Above-market compensation (discussed individually, depending on experience). What we offer: • Work with a live production product where AI avatars truly interact with users; • Opportunity to experiment with the latest models and architectures; • Minimum bureaucracy — maximum R&D; • Close collaboration with product and engineering teams; • Excellent chance to create a notable and innovative product; • Salary: negotiable, from 450,000 rubles.