Description:

The Staff Computer Vision Engineer will design and prototype vision-language architectures that integrate video streams with text for various use-cases involving sense-plan-act functionalities.
The role involves building end-to-end training pipelines on a small in-house H200 GPU cluster, managing distributed training, evaluation, and iteration processes.
Responsibilities include curating, augmenting, and managing large-scale video-text datasets, including generating synthetic captions to drive continuous model improvement.
The engineer will create automated testing and benchmarking suites for video QA, captioning, and instruction-following, using findings to refine models.
Collaboration with a lean team is essential to set technical direction and transition research code into deployable services.
The engineer will also work with the robotics team on the A part of the VLA.

Requirements:

Candidates should have 5-7 years of experience in deep learning, specifically in video computer vision (such as action recognition, temporal transformers, or video classification) or large-scale NLP/LLM fine-tuning.
Proven experience in large-model training on multi-GPU clusters using PyTorch, including distributed and mixed precision training, is required.
Strong skills in Python engineering, data pipelines, and maintaining experiment hygiene are necessary.
Candidates must be comfortable handling large video datasets (including IO, sampling, and augmentation) or large text corpora.
Effective communication skills in small teams are essential.

The position offers a competitive salary and stock options.
Employees can work remotely from anywhere in Europe.
There is an option to work in a new office located on the AI Campus in Berlin, collaborating with other innovative AI ventures.
Employees will be at the forefront of defining artificial intelligence applications in manufacturing.
The role provides hands-on experience in an AI-first software company.
The company promotes a supportive and inclusive culture that values diversity and encourages the advancement of underrepresented groups.
Employees will collaborate with a diverse team of over 10 nationalities on cutting-edge projects with real-world impact.
Networking opportunities with professionals and leaders in the field will be available, potentially opening doors to future career opportunities.
The company maintains a flat hierarchy, offers open 360° feedback, and provides flexible working hours.
There is a commitment to developing ethical AI software.