Specialization in Machine Learning with Google Cloud Platform
Proud to be the first Google Cloud Partner in Spain specialized in Machine Learning
Recently, Emergya has obtained the specialization in Machine Learning as Google Cloud Partners. It has been a long process, and it had its difficulties. In fact, we are the first Google Cloud partner who has achieved the Machine Learning specialization in Spain, and the fourth at European level so we are doubly happy and proud. Let's tell you a little about how the process has been, and what are our main conclusions after it.
What is the partner specialization in Machine Learning
As it is said from Google Cloud: Specializations match a customer’s need with a partner’s expertise in a specific service or solution area. By achieving specialization you signal to the market that you have gone through a rigorous technical assessment, employ certified technical professionals, and have demonstrated customer success in your area of specialization.
Currently, there are the following areas of specialization in Google Cloud Platform:
Development of cloud applications
The specialization in Machine Learning, in particular, is not easy to obtain. As Google says: "Partners who have achieved this specialization have done so because they have shown to have more than enough skills and experience in carrying out data exploration and preprocessing training, evaluation, and implementation of models, online prediction and knowledge. and experience using Google's machine learning APIs."
The Google Cloud Platform products included in this specialization are TensorFlow, Cloud Machine Learning Engine, Cloud Jobs API, Cloud Natural Language API, Cloud Speech API, Cloud Translation API, Cloud Video Intelligence API y Cloud Vision API.
What are the benefits of obtaining the specialization?
The benefit of obtaining the specialization in machine learning, apart from the obvious external recognition or appear as specialized partners in the Partner Directory, is to have gone through the process of specialization itself. This process has allowed our technical teams to test both their knowledge and their way of working, and grow and learn from the expert engineers in Machine Learning of Google Cloud in Mountain View, which has been a benefit not only for us but also for our clients and both current and future projects.
What are the requirements?
There are 4 requirements to be able to access the specialization:
- Having at least 4 employees who have the Data Engineer en Google Cloud
- Being able to provide at least 3 cases with clients where Machine Learning techniques on Google Cloud have been successfully applied
- Presenting a business plan with investment, return and growth objectives in case of obtaining the specialization.
- Passing the technical assessment
Each of the points above are validated by a different Google Cloud team, a specialist in a specific area. The most complicated of the four has undoubtedly been the technical evaluation carried out by Google Cloud engineers in Mountain View.
Technical capacity in Machine Learning
The technical evaluation required us to create 3 different demonstrations:
- Using the estimation API in TensorFlow to create a classification or regression model with structured data
- Implementing a customized estimator for a category of models that are currently not compatible with the estimation API
- Doing a demo with one of the Google Cloud Platform's Machine Learning APIs. In our case, we did it using the Vision API, Natural Language API, and Translation API.
All demos were created from real projects, which then benefited from feedback from the evaluation process. For each of the demos we had to create a GitHub repository that we shared with Google engineers, a pipeline for the ingestion and securitization of the data in the GCP storage, an end-point where you can test the developed models and a technical whitepaper that explained the main stages in the development, the why of certain technical design decisions and the issues that we had been dealing with along the way.
Security in Machine Learning projects
To finish, we had to explain in depth how we assure that in Machine Learning projects we address the associated security and privacy problems in an appropriate manner. For example, how do we ensure that the confidential training data stored in GCP is well protected, how do we carry out the processes of de-identification (masking, grouping, etc.) of data sets, etc ... Security and anonymity maintenance in the data provided by our customers is, if not the most important, one of our main concerns. And we take it really seriously.
Is the specialization worthwhile?
After having successfully passed through the specialization process, our conclusion is that carrying out this process is totally worthwhile. It is worth the end of the road, for the recognition that Google Cloud gives you as a specialist, but above all, it is worth the road itself. It is a process of learning and maturity at technical and process level, that will force you to investigate and get the most out of both the knowledge of the team and the Google Cloud Platform tools.
And of course, those who benefit the most are your own clients, who can count on teams of recognized professionals who have been successfully defending their work in front of some of the world's most outstanding machine learning engineers: those from Google Cloud.
In case you are interested in the subject and have any questions, or want to take advantage of our knowledge in Machine Learning and Google Cloud Platform, do not hesitate to contact us. We will be happy to assist you.