Research Scientist - (Senior/Principal)

techire ai; San Francisco (CA); Presencial; USD 180.000 - 260.000; Jornada completa

Descripción de la vacante

techire ai is seeking a Research Scientist / Principal Research Scientist specializing in vision-language models, video understanding, and post-training. You will conduct hands-on research within a small team and contribute to production-ready models.

The role covers training, evaluation, and deployment, with emphasis on long-context and egocentric video, SFT, RL, and efficient post-training techniques to boost real-world performance.

Formación

• Hands-on research in multimodal VLMs and long-form video understanding.
• Experience training transformer-based models and post-training approaches.
• Publications helpful; practical research impact valued.

Responsabilidades

• Train and improve vision-language and multimodal models on large-scale real-world video.
• Develop post-training pipelines using SFT, RL, and reward-based training.
• Advance long-context and egocentric video understanding with temporal reasoning.
• Design training objectives for temporal, spatial, and activity reasoning.
• Build evaluation frameworks for video understanding and real-world model performance.
• Explore distillation, quantisation, and model routing for production deployment.

Conocimientos

Deep learning

Computer vision

Multimodal learning

Video understanding

Transformer models

Reinforcement learning

Post-training

Descripción del empleo

Research Scientist / Principal Research Scientist - VLMs, Video Understanding & Post-Training

Interested in working on multimodal VLMs, long-form video understanding and post-training where the research has a direct path into production?

This is a hands-on research role within a small team building models for complex real-world video understanding using large-scale proprietary video data.

You'll work across training, post-training, evaluation and deployment, with a particular focus on long-context and egocentric video, temporal and spatial reasoning, and improving model capabilities through SFT, reinforcement learning, reward-based training and other post-training approaches.

Your work will include:

• Training and improving vision-language and multimodal models on large-scale real-world video.


• Developing post-training pipelines using SFT, RL, preference/reward-based training and parameter-efficient tuning approaches.


• Working on long-context and egocentric video understanding, temporal reasoning and memory.


• Designing new training objectives and experimentation strategies for temporal, spatial and activity reasoning.


• Building evaluation frameworks for video understanding, activity recognition and real-world model performance.


• Exploring distillation, quantisation and model routing for efficient production deployment.

At Principal level, you'll also help shape the research roadmap, set technical direction, lead major model initiatives and mentor other researchers.

You'll have a strong background in deep learning, ideally across computer vision, multimodal learning or video understanding, alongside hands-on experience training transformer-based models.

Experience with VLM post-training, reinforcement learning, reward-based training, long-form video or large-scale model development would be particularly relevant. A strong publication record is useful, but practical research impact is equally important.

The team has access to a large proprietary video dataset and dedicated compute, allowing researchers to test new training and post-training approaches on problems that are difficult to reproduce in a traditional lab setting.

If you're looking for a research role where you can push VLM and video models further through post-training, while still seeing the results used in production, we'd love to hear from you.