Machine Learning Intern Journal — aiXlr8
As the title indicates, this is the journal of a Machine Learning (ML) intern at the impactIA Foundation. I’ll be attempting to keep a weekly journal of my activities in the Foundation to keep track of my progress and leave a roadmap for the interns who come after me.
My third week at impactIA was mostly dedicated to an in-depth familiarisation of the Foundation’s aiXlr8 (accelerate) program. As briefly introduced last week, this project consists of the Foundation helping and advising SMEs (Small-Medium Enterprises) on how to integrate AI solutions into their businesses with a sustainable vision in mind. Let’s look at this in more depth. There are 5 main objectives in this program:
- Acculturation of Artificial Intelligence and its potentials for the company employees assigned to the program (1 day)
- AI Discovery Sprint (1 to 2 days) => Coaching for identifying a project that can benefit from AI optimisation
- AI Sprint (2 to 4 days) => Development of a SMIKE (technical prototype, an “alpha” version of the product)
- Mapping of skills to be developed and/or acquired within the company or to be outsourced to integrate AI projects
- Strategic development of a roadmap for the adoption of AI in the company
As discussed last week, one company we are working with are wanting to optimise the quality control of drill bits (drill heads) to check if there are any faults on the teeth (the tip of the drill head is made up of three teeth). Last Tuesday, we had a meeting with the CEO and leaders of the R&D department to present our first results in developing the SMIKE (technical prototype). They were impressed, but also eager to better understand the underlying principles of the system. We therefore agreed that I would come to their offices once a week to explain what is happening under the hood. This was an exciting yet daunting prospect, as the Foundation was giving me a lot of responsibility to interact with the clients without supervision. Yesterday was the first of these weekly meetings, and although I was quite nervous at the start, everything went fine. I explained the main outlines of our system without going too much in depth (as I’m still just familiarising myself with the actual code) and answered some of their questions. I was also given a tour of their factory, and it was impressive to see the production process. They have acquired a humanoid robot (a bust, two arms, a head with two eyes — but no legs) to perform laser engraving on small drill pieces, and it is quite disconcerting to see how easy it is to imbue it with anthropomorphic qualities. I wonder if it is a good idea to strive for more and more humanoid robots, or if we should develop non-humanoid robots (I’m looking at you, DAI).
This week, I started working on a second project with other clients, also related to quality control at the end of the production line. I met with the Director of Innovation on Monday for a tour of the factory, and was genuinely impressed by the automised machines working in perfect harmony (and also the amount of oil splashing everywhere to keep the process nicely lubricated!). This project is slightly less straight-forward in terms of gathering data, since we are looking to perform quality control on hexagonal pens, we need to find a way to capture all sides for image analysis. This is a problem we are currently working on solving.
Overall, this week has been a real step up in terms of responsibilities assigned to me, with two days of meetings and discussions with clients without any supervision. Although it felt quite intimidating at first, I’m very fortunate that the Foundation bestows this amount of trust in me after only a few weeks of working for them. Working on aiXlr8 programs is challenging and fun, since I get to use the knowledge and skills acquired during my degree for real-world problems — which as we were warned, are much, much messier than classroom problems! To be continued…