AI Jewelry Try-On API Development

Job ID: 39571198

Budget: ₹37,500 – ₹75,000 INR

Experience Level: Mid to Senior

Job Summary:

We are seeking a skilled Python Developer with prior experience in AI/ML projects to build and deploy
robust APIs for a virtual try-on system. The primary goal is to create an API layer around the
TryOnDiffusion library, which enables users to virtually try on clothes using a photo of themselves and a
garment image.

In addition, the developer will research and implement a similar solution for jewelry try-on, including
building APIs and optimizing performance.

Key Responsibilities:

· Integrate and build APIs using the TryOnDiffusion library to generate realistic virtual clothing try-ons.
· Deploy and optimize the service for efficient performance on the server.
· Handle input validation, image processing, and error handling in API endpoints.
· Research and develop a solution for jewelry try-on (e.g., earrings, necklaces) using AI or image
processing techniques.
· Optimize inference time and memory usage of the models.
· Maintain code quality, documentation, and version control.
· Collaborate with the frontend team to ensure proper API consumption.

Required Skills:
· Strong experience with Python and FastAPI or Flask for building RESTful APIs.
· Prior experience in machine learning or deep learning-based image generation or processing.
· Experience working with PyTorch and integrating ML models into production services.
· Familiarity with image handling libraries like Pillow, OpenCV, torchvision.
· Understanding of model optimization and GPU-based inference deployment.
· Comfortable working with Linux environments and Docker (optional but preferred).

Nice to Have:
· Prior work with virtual try-on, GANs, or diffusion models.
· Experience researching and implementing ML-based solutions for accessory try-on (e.g., jewelry).
· Understanding of modern AI tools for image synthesis, face detection, or landmark detection.
· Experience deploying models using TorchServe or similar model-serving frameworks.