Research Engineer

Bez empowers jewellers to use AI to their advantage. Our platform ingests sales data to generate original designs, manufacturing models (CAD/CAM compatible), and marketing videos in minutes. We work with leading diamond manufacturers and are scaling rapidly. Bez AI is supported by Macroscopic Ventures and Leap Year.

We’re hiring a Research Engineer (Generative AI, Vision) to turn cutting-edge ideas into shipped capabilities. You’ll prototype fast, validate with real data, and partner closely with engineering to deploy models that directly impact design tools, manufacturing specs, and marketing workflows.

What makes us unique:

• One User, One Model: Our agents fine-tune on each customer’s data to learn their brand aesthetics. With privacy-preserving post-training, no two users receive the same outputs.
• End-to-End AI: We don’t stop at images. Each design includes manufacturing-ready specs (3D render, technical drawing, BOM), and generations can be constrained by inventory (stone layouts, gold weights, chain styles).

What you’ll do:

• Own a practical research roadmap: ideate, prototype, evaluate, and ship.
• Build and refine multimodal generative models for images, video, and 3D-friendly outputs.
• Create editing/manipulation tools (inpainting, layout control, style transfer, text+image conditioning).
• Work with product and engineering to productionize models (latency, reliability, monitoring) and deliver to leading diamond manufacturers.

What success looks like:

• Weeks-to-impact: demonstrable improvements in generation quality, controllability, speed, or manufacturability.
• Clear experiments: metrics, ablations, and data-informed decisions.
• Production mindset: robust inference pipelines, reproducible training, and practical constraints baked in.

Required skills:

• Strong knowledge of computer vision and modern deep learning (diffusion/transformers, conditioning, control).
• Hands-on experience training, fine-tuning, and modifying generative models; comfort with dataset creation/curation.
• Solid coding skills: Python; experience with C/C++ or optimizing inference is a plus. Familiarity with tools like ComfyUI helpful.
• Strong CS fundamentals: algorithms, performance, debugging, and problem solving.

Nice to have:

• Multimodal experience (images, video, audio, text/speech).
• 3D vision and geometry (registration, articulated/non-rigid synthesis); familiarity with CAD/CAM outputs.
• Experience building creator/editing tools (masking, segmentation, tracking, temporal consistency).
• History of shipping research to product or measurable user impact.

Minimum qualifications:

• Demonstrated ability to deliver research engineering outcomes—prototypes that ship, models that improve product, or tools that make creators faster. Degree optional.

We’re an Equal Opportunity Employer and will consider qualified applicants with criminal histories in a manner consistent with applicable law.