Working Student – Data Labeling & Annotation (m/f/*)

Every model we ship is only as good as the data it learns from. As a Working Student in Data
Labeling & Annotation, you will sit directly inside our ML team and contribute to the quality of the
datasets that train the computer vision models at the heart of our product. This is not clickwork
at the edge of the organization: you will see how annotation decisions propagate into model
behaviour on real roads, and your judgment on ambiguous cases will shape how our systems
perceive the world. Your key tasks will include:

• Annotate image and video data for object detection, semantic segmentation, pose
  estimation, and tracking using our labeling tooling.
• Apply and refine our labeling guidelines, and help resolve ambiguous or edge-case scenes
  together with our machine learning engineers.
• Perform quality assurance on annotated batches, including review passes, consistency
  checks, and correction of systematic errors.
• Curate and triage raw footage: identify relevant sequences, flag failure cases from model
  predictions, and prioritize the data that closes real gaps.
• Support dataset bookkeeping such as class distributions, coverage statistics, and
  versioning of annotation batches.
• Handle all data strictly within our privacy-by-design processes and in full compliance with
  the GDPR.

Core Competencies: Precision & Visual Judgment

• Attention to Detail: A high tolerance for meticulous work and the discipline to stay
  consistent across hundreds of similar frames.
• Visual Reasoning: The ability to interpret complex traffic scenes – occlusion, motion blur,
  poor lighting, unusual objects – and to make defensible decisions about them.
• Guideline Thinking: The willingness to question a labeling rule when it produces
  nonsense, and to propose a clearer one instead of silently working around it.

Technical Stack & Tools

• Annotation Tooling: Comfort working with tools such as CVAT, Label Studio,
  Segments.ai, or similar – or the confidence to learn one quickly.
• Scripting: Basic Python or scripting skills for simple data handling and sanity checks are a
  strong plus.
• Working Environment: Familiarity with Git, spreadsheets, and structured file organization.
• Domain Exposure: Prior contact with machine learning, computer vision, or robotics
  through coursework, projects, or curiosity is a plus, not a requirement.

Formal Requirements

• Enrollment: Enrolled as a full-time student at a university or university of applied sciences
  in Germany for the duration of the role, in any field of study.
• Availability: 15 to 20 hours per week during the semester, with the option to work more
  during semester breaks.
• Location: Based in or around Berlin, with the ability to work on-site at least part of the
  week.
• Language: Excellent verbal and written communication skills in English; German is a plus.

Mindset

• Comfort with ambiguity, a problem-solving mindset, and adaptability to rapid change.
• A strong sense of responsibility – we will rely on your batches being right.
• Reliability and clear communication about your availability around lectures and exams.
• Genuine interest in how AI products are actually built, beyond what a lecture can teach
  you.
  We don't expect anyone to check all these boxes mentioned above! If a few of these points
  apply to you, we want to talk!

Deep insights into an AI startup: a seat inside the ML team of a venture-backed deep
tech company. You will sit in on our technical discussions, see how a model travels from
dataset to on-device deployment, and follow the product and commercial decisions that
surround it. Very few students get to observe the full loop from raw data to shipped product
at this proximity.

• Direct mentorship from senior machine learning engineers, and honest feedback on your
  work.
• Flexible working hours that adapt to your lecture and exam schedule.
• A realistic path into a broader role, a thesis project, or a permanent position as our
  company scales.
• A diverse and inclusive work environment as well as a flat organizational structure, fast
  decision making within the team, disagree & commit.
• We foster a culture where we all learn from each other and value new ideas.
• We hold fun team events throughout the year.
• We contribute to your monthly public transport ticket.
• Free drinks, snacks, and coffee

Please send your application and CV to jobs@peregrine.ai. Please also let us know your field of
study, your current semester, and your weekly availability.
Contact details
Jorit Schmelzle
CPO & Founder
Peregrine Technologies GmbH
Saarstr. 20a
12161 Berlin
jobs@peregrine.ai