Research Scientist, Vision-Language & Multimodal AI

mecka; New York (NY); Sur place; USD 160 000 - 230 000; Plein temps

Résumé du poste

Mecka AI is seeking an experienced Research Scientist to advance vision-language and multimodal foundation models. You will tackle challenges across model training, post-training, reinforcement learning, and multimodal reasoning, designing experiments and deploying scalable solutions.

You will work on data curation, evaluation, and benchmarking to push the boundaries of what multimodal AI systems can understand and accomplish.

Qualifications

• Hands-on experience researching, training, fine-tuning, or post-training large-scale foundation models.
• Strong understanding of deep learning architectures, optimization techniques, and model training methodologies.
• Experience with supervised fine-tuning (SFT), preference optimization, or reinforcement learning approaches such as DPO, GRPO, PPO, or RLHF.
• Proficiency in Python and PyTorch, JAX, or comparable deep learning frameworks.
• Experience designing, executing, and analyzing large-scale machine learning experiments.
• Familiarity with distributed training systems and GPU-based model development.
• Strong research instincts, analytical problem-solving skills, and the ability to independently drive complex technical projects.
• Experience working with open‑weight multimodal foundation models.
• Research publications on top-tier journals/conferences, open-source contributions, or experience within leading AI research organizations.

Responsabilités

• Multimodal Model Research: Research, develop, and improve vision-language models and multimodal foundation-model architectures.
• Post-Training & Reinforcement Learning: Design and implement advanced training strategies, including supervised fine-tuning, preference optimization, and reinforcement learning, to improve model capabilities.
• Model Performance & Optimization: Investigate approaches to improve reasoning, generalization, reliability, and training efficiency.
• Data Quality & Training: Develop strategies for data curation, filtering, and synthetic data generation to improve training outcomes.
• Evaluation & Benchmarking: Design experiments, evaluation frameworks, and benchmarks to measure model performance, identify failure modes, and guide research priorities.
• Large-Scale Experimentation: Work with distributed training infrastructure and large-scale datasets to support research and model development.
• Research Innovation: Explore emerging approaches in multimodal learning, model alignment, and foundation-model training.

Connaissances

Foundation models

Deep learning

Reinforcement learning

Python

Open-weight multimodal models

Publications / Open-source

Outils

PyTorch

JAX

TensorFlow

CUDA

Description du poste

Mecka AI is seeking an experienced Research Scientist to advance vision-language and multimodal foundation models. You will tackle challenges across model training, post-training, reinforcement learning, and multimodal reasoning, designing experiments and deploying scalable solutions.

You will work on data curation, evaluation, and benchmarking to push the boundaries of what multimodal AI systems can understand and accomplish.