Who are we? :)
Microblink is a research and development company with a mission to simplify data entry using the camera input. Using advanced machine learning methods, we develop state-of-the-art data extraction technology, linking modern apps with the physical world.
How we started...
In 2011 we started as a classic computer vision company specialized in real-time performance on mobile devices. Back then we were developing highly optimized computer vision algorithms written in C++, but in 2015 we encountered a problem that needed a different approach. We made our debut in 2016 with what we believe was the first handwritten math expression recognizer that worked in real-time on a mobile device. It was our first successful application of deep machine learning that powered tens of millions of devices. But more importantly, this started a 5-year transformation period for Microblink, where we invested heavily into our teams and infrastructure to enable a real data-driven development process.
Machine learning in Microblink
We have several products in production, powered by 15 machine learning models constantly being updated on new data and running in real-time on more than 200 million end-user devices. Production scale machine learning of this magnitude has its many challenges that our experienced team handles on a daily basis, but at least 20% of their time is focused on what we call future projects. These projects range from models that power new features and products to independent research projects that we may try to publish at a conference. Regardless of being published, every machine learning engineer in Microblink is sent to one of the top conferences in the field each year. Whether it’s NeurIPS in Los Angeles, ICCV in Seoul or BMVC in Cardiff, our engineers will be there, discussing machine learning with the strongest teams in the world.
Teamwork that enables focus
Jokes about machine learning engineers spending 80% of their time on cleaning data don’t apply in Microblink, we have teams that handle this by producing millions of annotations with highest possible accuracy.
Production training, evaluation and deployment is powered by our proprietary systems and development teams that aim to reduce manual overhead to the bare minimum and increase processing speeds to the absolute maximum.
This allows our machine learning engineers in Microblink to focus on machine learning.
We are looking for machine learning engineers to join our team on both computer vision and text analysis student positions!
Apply and become our Machine Learning Engineer student (f/m)
As a Machine Learning Engineer in Microblink, you will:
- Develop and update Microblink’s machine learning models for production
- Research and develop new machine learning models that power Microblink’s future products and features
- Perform independent machine learning research and potentially publish your work at a conference
- Work closely with Microblink’s experienced software engineers on our products and/or infrastructure
- Learn, improve and develop your understanding of machine learning
You are the right person for us if you:
- Have a solid understanding of machine learning in the domain of computer vision or text analysis
- Can write sustainable, modular and efficient code in Python
- Have experience with machine learning tools such as TensorFlow or PyTorch
- Can work well in a team, but are also capable of doing tasks on your own
- Can equally well do tasks with creative focus as well as those with repetitive, more operational focus
- Nurture proactive and responsible approach to work
- Eager to learn about new trends in your field
- Have great English language skills (both written and spoken)
An additional plus (but not must) is if you:
- Have experience in C++ development
- Have experience with DevOps technologies
What are we offering?
- Flexible working hours/schedule to support your focus on University studies and obligations
- Learning opportunities through
- Direct mentorship - when you join our team, you’ll get a dedicated mentor, who will be your counselor, consultant and cheerleader
- Knowledge sharing - we make time for/dedicate 20% of our working hours for education
- Educations and conferences - we are regularly attending some of the biggest machine learning conferences worldwide such as NeurIPS, ICCV, and BMVC
- Dedicated budget for professional development and education
- Internal library that keeps growing, based on our needs and interests
- Online courses
- Opportunity for full-time employment upon graduation
- A culture that recognizes and rewards success, and is not afraid to try, fail and learn from the mistakes
- Last, but not least, our main strength is teamwork and a friendly atmosphere.