Build multimodal intelligence for Palona’s physical-world AI using visual grounding, temporal reasoning, and structured prediction from video.

This AI Research Engineer role focuses on vision-language modeling and computer vision for busy restaurant environments. You will develop approaches for changing scenes, incomplete information, and difficult visual conditions. The work spans research questions, model adaptation, dataset creation, evaluation, and production collaboration.

Key Responsibilities:

- Develop VLM approaches for events and activities across video
- Fine-tune models for grounding and structured prediction
- Create representative image and video datasets
- Build benchmarks for temporal consistency and perception quality




- Translate research advances into practical product capabilities

Requirements:

- Three-plus years of applied or research experience
- Depth in video understanding, grounding, tracking, or multimodal reasoning
- Hands-on model training beyond basic API integration
- Robust Python and PyTorch or equivalent framework experience
- Experience analyzing failures and using ablations

Bring research depth in multimodal learning, visual grounding, and video understanding to Palona’s restaurant-focused AI products.

📌 Multimodal Vision Research Engineer (Toronto)
🏢 Palona AI
📍 Toronto