Research Scientist, Multimodal Vision–Language Foundations

Mecka; New York (NY); On-site; USD 185,000 - 210,000; Full time

Job summary

Mecka, based in New York, is seeking an experienced Research Scientist to advance multimodal foundation models. You’ll work across model training, post-training, evaluation, and optimization to push the boundaries of vision-language systems.

The role emphasizes hands-on experimentation, scaling large experiments, and translating breakthroughs into real-world capabilities with a focus on vision-language models, reinforcement learning, and data pipelines.

Qualifications

• Hands-on experience researching, training, fine-tuning, or post-training large-scale foundation models.
• Strong understanding of deep learning architectures, optimization techniques, and modern model training methodologies.
• Practical experience with supervised fine-tuning, preference optimization, or reinforcement learning approaches such as DPO, GRPO, PPO, or RLHF.
• Strong proficiency in Python and ML frameworks (PyTorch, JAX).
• Experience designing, executing, and analyzing large-scale machine learning experiments.

Responsibilities

• 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 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 and synthetic data generation to improve training outcomes.
• Evaluation & Benchmarking: Design experiments, evaluation frameworks, and benchmarks to measure model performance.
• Large-Scale Experimentation: Work with distributed training infrastructure and large-scale datasets.
• Research Innovation: Explore emerging approaches in multimodal learning and foundation-model training.

Skills

Python

Deep Learning

Multimodal Models

Research Experience

Distributed Training

RLHF / PPO / DPO

Tools

PyTorch

JAX

Job description

Mecka, based in New York, is seeking an experienced Research Scientist to advance multimodal foundation models. You’ll work across model training, post-training, evaluation, and optimization to push the boundaries of vision-language systems.

The role emphasizes hands-on experimentation, scaling large experiments, and translating breakthroughs into real-world capabilities with a focus on vision-language models, reinforcement learning, and data pipelines.