ML инженер

Higgsfield AI develops a generative AI platform focused on video and image generation models.

• Design and implement end-to-end training pipelines for state-of-the-art models in computer vision, including video and image generative models;
• Develop complex data pipelines to transform raw inputs such as images, video, text, and audio into high-quality, annotated datasets that power model training and evaluation;
• Work tightly with scrapping and annotation teams;
• Optimize models for speed, scalability, and efficiency;
• Develop tools and frameworks to accelerate training, evaluation, and deployment of large models;
• Collaborate with product and design teams;
• Beat other models on public benchmarks such as LMArena and Artificial Analysis.

• Proven track record of training and deploying ML models into production;
• Strong skills in model training, optimization, and evaluation, with hands-on experience in distributed training and multi-GPU systems;
• Knowledge of modern ML research trends, frameworks, and tooling, including PyTorch, Hugging Face, DeepSpeed, Accelerate, and Ray;
• Experience with evaluation of complex ML systems;
• Willingness to work in a fast-paced, high-intensity startup environment;
• Excellent communication skills and ability to collaborate across research, engineering, and product teams;
• Nice to have: Experience with large-scale vision, especially diffusion models, NLP, or multimodal systems.

• Competitive salary in USD;
• High growth potential.