About the role
Responsibilities:
- Build and deploy cutting-edge vision-language pipelines (e. g., Stable Diffusion, LoRA, CLIP, BLIP) for personalized and authentic user profile visuals.
- Fine-tune LoRA models to reflect a user's best version while maintaining identity integrity.
- Design scalable APIs for image generation, editing, and personalization.
- Work with UX/design to integrate image-generation flows that feel intuitive and delightful.
- Optimize models for mobile performance, fast inference, and cost efficiency.
- Collaborate closely with product, backend, and ML researchers to productionize ideas.
Requirements:
- We are looking for a hands-on applied AI/ML engineer with expertise in generative vision models like Stable Diffusion, and some track record of building real-time, personalized generative AI experiences.
- Experience with computer vision, generative models, or deep learning.
- Experience working with Stable Diffusion, LoRA training, and prompt engineering.
- Exposure to different vision-language architectures like Flux, CLIP, BLIP, etc.
- Proficient in PyTorch or equivalent ML framework.
- Strong understanding of image pipelines, dataset curation, and captioning.
- Experience deploying ML models on cloud-based services ( at minimum Replicate, Hugging Face Spaces, and preferably on GPUs).
- Comfortable writing production-ready Python, working with versioned models and real-time inference stacks.
- Passionate about building beautiful, human-centered AI experiences.
- Prior experience with personalization, avatars, or face-preserving generative models.
- Contributed to open-source vision models or fine-tuning libraries.
- Experience in a startup or fast-paced product-focused team.