Development collaboration with Ambolt AI

AGCO Innovation Center Randers is an innovation center within the Global AGCO group which explores and delivers new ideas and solutions in software, hardware and mechanics.

AGCO Innovation Center Randers constantly has projects on the go, both large and small, and this is where Ambolt AI comes in, offering highly specialised consultancy work within AI technologies. Without being able to say what the end product is, we can safely say that we help AGCO to create ground-breaking solutions that will in turn create immense value for the end users!

Website

The collaboration has come about to support AGCO Innovation Center Randers’ core competencies and to bring in specialised knowledge, particularly in Computer Vision, for various projects. The flexibility inherent in having an external partner with specialised knowledge, not least up-to-date knowledge of state-of-the-art technologies, is also achieved. In addition to Computer Vision, the projects involve work with neural networks, programming in C++, ROS and Python.

The collaboration with Ambolt AI has run like clockwork, and the consultants have been a natural part of our development team from day one, adding value to the projects. Not only have they brought new methods and ideas to the table, but they have also been incredibly proactive in their collaboration with the rest of the team. Proactive collaboration is particularly important to us as it allows us to avoid processes and maintain our agile way of working.
Kenneth Düring Jensen. Mgr, Research & Advanced Engineering, Gold Harvesting

Build – Measure – Learn

The agile culture at Innovation Center Randers perhaps explains the ease with which Ambolt AI’s consultants have slipped into the various teams and projects. We share the same way of thinking when it comes to new development, where software is developed based on “The Agile Manifesto” and “Build-Measure-Learn”.

This agile approach to projects ensures that everyone is aware of the goal and the distribution of resources, making for an effective collaboration that can function on both a physical and a virtual level.

Start small

One of the first projects Ambolt AI was involved with was the testing of an idea.

Was the idea even feasible?

And was there a basis for starting a project?

Here, we started with completely raw data and set about testing the project’s hypothesis using various techniques to achieve Proof of Concept. The conclusion in this case was that yes, with more data the project is doable. This paves the way for a project that will make a certain type of machine more efficient and significantly improve the quality.

Scale fast

After Proof of Concept, Ambolt AI has also become involved in the next step in the process, which is to initiate the project itself and build the infrastructure. We are building the project’s pipeline, starting from how data is collected, through data annotation, and right up until it is used for training models in a Computer Vision/Deep learning solution.

The collaboration on this project involves weekly status meetings, with praise for both quality and the pace at which the project is progressing. Results have been achieved in record time.

Ongoing projects

Contact Ambolt AI

Not having been involved in a large project from the outset does not rule out the possibility of adding value. On the contrary. Ambolt AI joined a large project a couple of years into the project phase.

After a brief start-up phase with close collaboration at AGCO Innovation Center Randers to get to know the team and become familiar with the project, Ambolt AI moved on to status meetings and workshops. The regular follow-up meetings and workshops help to answer any questions and clarify the direction of the project – and not least to create a sense of community around it, which is especially valuable when external consultants are part of the team.

Again, it is Ambolt AI’s down-to-earth approach and expertise that makes it possible to add value to the project.