Senior Computer Vision Scientist/Engineer (Multiple positions)

Permanent employee, Full-time · Berlin/Remote

What is Yaak?
Yaak is developing a driver coaching platform (SafetyOS) that prioritizes safety. Partnering with our driving school customers around the world, we codify into SafetyOS their decades of experience in scoring, detecting, explaining and correcting human driving. We actually don't do anything with yaks.
Why human driving?
The year 2007 heralded new possibilities in autonomous driving. The excitement wasn't unfounded of course. We had solved driving in the rugged terrain of the Mojave Desert, with no internet connection and without supervision. Full Self-Driving seemed like a few years away. A decade, a deep learning revolution, and many $B later we are still waiting for it, if/when it's unaffordable.

At Yaak we asked ourselves how to go about building AI that is provably safe. How can we bake trust into it from grounds up? Instead of building a classic AV stack, why not build a safety-first driving platform that can score, detect, explain and correct human driving with high precision before it's allowed on public roads.
The Role
SafetyOS — our flagship product built for and with driving schools, aims to solve this. With SafetyOS we codify decades of driving instructor's experience in scoring, detecting, explaining and correcting human driving. Leveraging state-of-the-art camera sensors, compute, vehicle telemetry coupled with driving instructor annotations and OSM data, you'd be tasked with building transparent, efficient, high precision models driver scoring models. Within the AI team you'd tackle one of the biggest challenges in Computer Vision and Machine Learning, with freedom to throw away old ideas and start from first principle. Be that building an efficient containerized distributing training infrastructure, bringing in latest and greatest mlops tools and knowledge to the team or designing ML models inspired from OSS or from scratch.
You won't see hard requirements here for certain degrees or exact years of experience because we know everyone has their own unique journey. If you're brilliant at your craft, and are either an experienced collaborator or motivated to learn and excel at it, we'll know it when we see it.

You have:
Big plus for:
  • Fair understanding of Generative modeling with VAE/VQ-VAE/GANs/Flows
  • Experience with moving research ideas into production with stakeholders
  • Experience with self-supervised learning on large scale visual data (Image/Videos)
  • Fair understanding of sensor calibration pipelines & depth estimation
  • Prior knowledge of model compression for compute constrained devices
Bonus points for:
  • Competitive compensation package with employee stock options program
  • Yearly public transport ticket and mobility bonus
  • Flexible work-from-home policy (even when the world is pandemic-free)
  • Quarterly off-site somewhere in the world, to regroup with the team and plan
  • Work in an international team full of talented, chill and cooky people
  • Visa and relocation support

Distributed-first perks

  • Free access to corporate apartment in Berlin (Yaak HQ)
  • Company credit card for managing regular travel to Yaak HQ 
Diversity in Tech
Yaak is committed to fostering diversity and inclusion. We want you on our team — no matter your ethnicity, gender, religion, sexual orientation, age, disability, or if you’re currently parenting the next world leader. It's our belief that the best work gets done when one feels appreciated, supported and are their most authentic self.
We are looking forward to hearing from you!
Thank you for considering a career at Yaak Technologies. Please fill out the following form. In case you are experiencing problems with the document upload, mail your documents to .

Please upload your CV and any other supporting documents

Click to select multiple files or use drag-and-drop
Click to select multiple files or use drag-and-drop

Uploading document. Please wait.