Dr Andrea Isoni

All data is initially anonymized or redacted for general team use. Techniques like data masking, tokenization, and automated redaction remove sensitive information. This protected version is used for most work and processing. Only a select few team members have access to the original, unaltered data. Strict access controls, including role-based permissions and multi-factor authentication, safeguard the original version. A secure process is in place for authorized personnel to access original data when absolutely necessary. Regular audits and employee training reinforce this two-tier system, balancing data utility with strong privacy protection.

Writing a newsletter or blog post can be an example of establishing yourself as a thought leader. The trick is to choose a subject within the AI space that both: -you are obsessed about -there is an audience for it You need to be obsessed because you need to constantly study and write about that subject. You cannot expect results if you write once a year. Of course, the subject you write also has a good audience. By good I mean an audience with a large number of people, possibly growing and somehow helpful for your career. To get the right combination of all the elements above it is not easy. It is relatively easy to find something you like to write about, it is easy for subjects that have a market but it is not easy to combine them.

Wearable technologies are coming in the next decade or so. AI pins are just an example, but there are also tech glasses, smart shoes etc. Even today you can use machine learning systems to try online clothes, like a 3D fitting room or help in sales assistance (like an assistant that helps the person trying the clothes online: usual suggestions like 'try blue instead red etc.'). All these systems by definition will rely evenly on machine learning algorithms.

If you really believe your ideas are good but people do not get it, well you should use social networks to your advantage like Linkedin. What I mean is to post your ideas around and explain that to different audiences to gain support/feedback. Here are the advantages: -you will learn how to explain things outside your usual colleagues or peers -if the idea is right and people around get it, you know now who is your 'target audience' for support -if the idea is right and people around DONT get it, well later on you can show you have said it first. Especially in AI, where things change quickly it is a good habit to reach a diverse audience for feedback/support.

You need to be able to explain difficult machine learning concepts in simple terms. The more you can do that, the more you show both you grasp the concepts and you can interact with business stakeholders. Ideally you should be able to explain in plain English concepts most known only by the mathematical definition: as an extreme example, what is a fully connected layer in a neural network. In a normal interview you definitely need to know how to translate paper symbols into code: for example if in the code the variable is X_train, you need to find yourself what's the equivalent in a formula in a research paper. It seems simple, but many in an interview environment may fail to 'translate' code names into papers' formulas names.

In any case an internship is your way to find a real job after. So having more time is not necessarily a bad thing. It is at that point up to you to make sure the internship is an 'ml internship': expanding skill does not mean only technical skills but to own the project fully. If necessary, 'invent' a project within the scope of the initial internship that: 1. use the data, or initial scope of the internship towards an ml objective 2. if you do not have the data, generate them (and learn how to do it) 3. it has a quantifiable metrics 4. ideally it solves a business case 5.ideally it uses latest research or github/huggingface repositories available If you do that, the internship will have the desired effect: help you find a job.