ML Engineer - Personalization & Recommendation Systems

Krea AI Krea AI · Multimodal · San Francisco, CA · Research

ML Engineer to architect and build Krea's personalization and recommendation systems from scratch, focusing on user taste, content curation, and adapting generative models to individual aesthetics. This role involves designing algorithms, building curated feeds, and contributing to personalized image generation research, taking systems from research to production.

What you'd actually do

  1. Architect and build Krea’s personalization and recommendation stack from the ground up, owning the technical direction end to end.
  2. Design algorithms to model user preference and taste, enabling Krea’s models to adapt to individual styles and aesthetics.
  3. Build high-quality, curated feeds that balance exploration, personalization, and aesthetic coherence.
  4. Work directly with the our model and research team to co-design personalization mechanisms that influence how our generative models learn, adapt, and express style.
  5. Contribute to personalized image generation research, with a focus on style, taste and subjective quality.

Skills

Required

  • Strong experience building recommendation systems or personalized feeds from scratch
  • Proven ability to design and ship high-quality curated content experiences
  • Experience working with media-based personalization (image, video preferred; music or other modalities also welcome)
  • Solid foundations in machine learning, representation learning, and modern deep learning techniques
  • Strong Python skills and experience with ML frameworks such as PyTorch or JAX.
  • Ability to operate independently, make architectural decisions, and own complex systems end to end

Nice to have

  • Experience with large-scale data systems and production ML infrastructure
  • Prior work on or familiarity with diffusion models or generative image systems
  • Background in embeddings, similarity search, ranking, or aesthetic evaluation
  • Interest in creative tools, art, design, or generative media

What the JD emphasized

  • building recommendation systems or personalized feeds from scratch
  • design and ship high-quality curated content experiences
  • media-based personalization (image, video preferred; music or other modalities also welcome)
  • operate independently, make architectural decisions, and own complex systems end to end

Other signals

  • building personalization and recommendation systems from scratch
  • model user preference and taste
  • curated feeds
  • personalized image generation research
  • take systems from research and prototyping through production