My Internship Experience at a Space Engineering Practice
Image Credit: https://www.scitecheuropa.eu/the-new-frontier-for-space-exploration/97776/
Craft Prospect, a growing space engineering practice in the heart of Scotland’s thriving CubeSat sector, has been my professional home for the last 6 months. As part of my PhD program, I was fortunate enough to have the opportunity to do a placement in industry. One of the first company’s on my radar was Craft Prospect. Founded in 2017, the company’s primary goal is to use a combination of small satellites (CubeSats), lasers, sophisticated software and quantum technology to provide augmented future encryption services. Much effort has recently been spent by Craft Prospect to exploit new developments made in the now burgeoning field of machine learning for future space missions.
Machine learning has seen a resurgence in popularity across a vast array of sectors including finance, tech, agriculture and academia. It is used in just about every aspect of our lives such as automatic tagging of photos on Facebook, Spotify music suggestions, credit card fraud detection, spotting diseased crops and Google Maps route predictions (to only name a few).
But what makes machine learning so useful for so many people and why are we using it at Craft Prospect?
Answer: It’s fast and incredibly reliable at solving well-defined problems given a large amount of data. Luckily, we live in an age where new data is being produced at incredible rates all the time (sometimes even terabytes per second!) and is often provided open-source to the general public. Because of this, and many other factors, machine learning was an obvious avenue to explore for our work.
Over the course of the past 6-months I had the opportunity to work on several projects where we applied novel machine learning techniques to optimise future CubeSat space missions. Because of the structure of the company, I was given lots of autonomy to pursue my projects in what I thought might be the most effective way possible. As part of my internship, half my time was spent at the company office and the other half at the University of Glasgow where I am pursuing a PhD in gravitational wave astrophysics.
Projects would typically span 3-month periods in distinct phases: conception, development and implementation.
Conception:
- Other engineers and I would go around spit-balling ideas about potential exciting applications of machine learning to our own real-world problems. We’d then hone-in on the most promising idea that could make the greatest impact within a reasonable amount of development time.
Development:
- During development, we then take a proposed concept and turn that concept into a workable solution. This might involve, but would not be limited to, background research, further discussion with other team members, writing up an initial bare-bones framework for the code, familiarisation with potentially useful software packages and finally filling that bare-bones framework with code.
Implementation:
- Once code has been written and has passed diagnostic tests, we would proceed to debug and tidy it up to make it as user-friendly as possible. If the results are exceptional enough, we then repackage the outcome of the project in the form of a grant proposal to get more funding or a journal submission.
There are a great many fantastic machine learning resources freely available to the general public I personally have used both while at Craft Prospect and during my own PhD studies. A good starting point for understanding the underlying concepts is the Machine Learning course by Andrew Ng. Once you’ve come to grips with the background, the Keras high-level python API for machine learning is an excellent package for actually implementing your ideas! StackOverflow will be your best friend for any and all general coding debugging issues.
All in all, it was a great experience getting to work as part of the Craft Prospect team alongside so many talented, passionate people. It was a great experience to see how every person in the team works as an individual or in small teams to tackle different aspects of a mission. I ended up learning quite a lot in a short amount of time and had a blast doing it. If you’d like to find out more about Craft Prospect, feel free to check out the following website for any information on past, current and future missions.
Thank you! If you have any questions or just want to connect, feel free to contact me on email at h.gabbard.1@research.gla.ac.uk, or LinkedIn.
