Meet Mackenzie. Algorithm writer, ‘inconceivable’ designer and e-scooter rider.

Mackenzie Dreese breaks the mold. She is making people and the machines she works with smarter about the possibilities of the future of energy – and commutes to work with her e-scooter.

Mackenzie, could you please tell us a little bit about your background and journey so far?  

I started in the Oilfield Equipment product company, the subsea drilling group, as a design engineer for blow-out preventers. Then I moved to the Energy Innovation Center (EIC), in Oklahoma City, as part of the Baker Hughes rotational program I’m in, called ASPIRE.  My background is mechanical engineering and physics from an undergraduate perspective, and when I came to the EIC, I just fell in love with the culture. The research forward mindset the staff there has, how everyone in the building has advanced degrees and PhDs, how much a part of their lives research is…it is so inspiring!

What do you work on and how have you been able to bring some of your experience to your role?

I started working on a lot of data science stuff: machine learning and computer vision, within the reservoir innovation group. On the core analysis team, I was writing computer vision algorithms to identify lithotypes, fractures and other features in the core. After that rotation, I got to do a lot of cool things: for example, I presented at a conference in France, my first time traveling abroad for work.

Now, I am at the Additive Manufacturing Technology Center (AMTC) in Houston for my last rotation, for 1 year. I actually pitched the idea to go there. Usually you get placed by your program manager, but I saw the gap between using machine learning and AI out of AMTC. So, I set up a meeting with the AMTC leadership team and I told them how much I thought machine learning could be adding to this group. It was exciting.

Once I got here, I was working with a CT scanner again.

At the EIC, we were scanning core with large medical-scale scanners, like the one used when you go to a doctor. And here in Houston, we were just starting to CT scan parts for additive manufacturing. Scanning fully finished printed parts to identify their geometries, if they have any cracks, voids, delaminations or defects of any kind.

I had an algorithm that I wrote during my time in Oklahoma City: so, I translated it immediately to this group, and with only some minor retraining of the algorithm, using this new additive data instead of the original core data. It has been working to identify fractures, cracks, delaminations in the printed parts. A true eureka moment: using the exact same algorithm from the other group – and it was a fairly easy translation. to me that speaks a lot to the power being able to cross-pollenate knowledge. It was very cool! Now, I am working on all kinds of things at the AMTC, machine learning and AI – a lot of it is built around CT.

I have also another project where we are taking features from parts provided to us by some of our customers. Using an algorithm of predictive neural network, we can help determine whether a part is suitable for additive manufacturing, based on all the parts that we have printed in the past, as my training data set. It’s been fun to use it for both internal and external customers.

What’s exciting about Additive technology?

There are a lot of things that are exciting about additive.

We are so obviously determined to push the envelope on what you can do with additive. From a design perspective, from an inspection perspective, we really are trying to be cutting edge in everything. And we are beginning to offer services that are really ground- breaking, in my opinion. We are now offering design-as-a-service, inspection-as-a-service and digital qualification.

I think the coolest thing about additive is the doors it opens up from a design perspective. As a mechanical engineer, we are so used to creating parts in a very traditional way, using very traditional manufacturing techniques. This group likes to use the phrase ‘inconceivable design’. 

Another member of my group is working on a simulation and design project basically doing machine learning-aided/ AI-aided design. They are working with an external client on this and it has been really interesting to see prototypes come to life: the simulations and algorithms are coming up with designs which you could never see humans come up with. They are so intricate: with spirals and lattices.

This is a really exciting direction to be going in, just to imagine the types of designs additive is capable of creating, simply because there are very different boundary conditions to what you can and can’t design with this approach.

How do you think this technology can impact oil and gas operations in the future?

When people first think about additive manufacturing, they jump to Aviation, just because it’s the biggest sphere, where additive was active first and other industries got on that train very early. When you think of additive manufactured parts for aircrafts, you automatically think: ‘we have to make this part as light as possible’.

Well, it is not necessarily the case for oil and gas. The benefits of additive for oil and gas applications are very different. We are putting parts in these extremely High Pressure/High Temperature environments. We are designing many parts for additive that can bring real reliability improvements, not just in the sense that we are taking away weight from something. There are examples where we have re-engineered a part for additive to have a much lower failure rate than its traditionally manufactured counterparts. In oil and gas, this is obviously one of the biggest cost savings: reducing downtime.

In our sector, we truly are writing the rule book on corrosion-resistant additive right now: our teams are doing studies on parts we printed, partnering with materials organizations such as NACE. Working together on formulating what can and can’t be a material for additive, in these high-end corrosive/HP/HT environments where additive can really play a significant role. This is so important, especially when we get to the level of designing at the micro-structure level.

What do you think is a common misconception about the work you do when you talk to friends/ colleagues?

Obviously, yes, there are misconceptions about Additive.

Many assume that additive makes a part more expensive than traditional manufacturing; that is not necessarily the case. When you consider how many ways there are to improve a product with additive, it can be in the ‘literal’ cost saving on a part, where the part becomes much cheaper to make.

However, I think there are even more misconceptions about machine learning and AI itself. People don’t really understand what machine learning is and how it works, and I have been really lucky to date to work with many who have been very receptive to my explanation of what I do every day.

My Dad for example, he doesn’t really understand what I do,- but he always refers to it as the movie ‘The Terminator’ when I talk about machine learning and AI.

And I have to say: ‘No Dad, we are not exactly to the point where we are about to start a robot revolution by accident but we are able to teach machines to identify certain things.’

One of the most famous examples of where we are in AI right now is that algorithms can currently  as advanced as they get  identify that ‘this is a dog’ vs ‘this picture is a cat’…So that algorithm is certainly not really about to trigger some sort of a robot uprising anytime soon!

The power is that you can translate it to industry and as a result you can automate tasks that, honestly, engineers should not to be doing in the first place. Having highly skilled materials scientists look through thousands of images of a CT scanned additive part just to see if there is a defect in it can take many hours. If this can be done by a computer instead: it will be much more efficient for everyone, and lets us focus on the real hard science, the real problems. So that’s what I tend to describe it as.

And my work is very data driven. That’s the other big misconception: that machine learning and AI can solve anything, that it is this big beautiful solution to everything. But it is not some grand unifying equation. Anything that I write, or that anyone writes is going to be only as good as the data they can provide to it. They are the big things.

To close: why did you choose to work in this industry, when you could probably work in any of the big tech companies?

I have always been fascinated by energy, but like many young people, I am very climate change conscious, very passionate about protecting our oceans, and preventing spills. And I also recognize that hydrocarbons/fossil fuels can’t go anywhere anytime soon, and while this is the case, I believe it’s important to make it as safe as possible.

And to be with a company that brands itself as a part of the energy transition is a really big deal, having that larger perspective was very enlightening and eye opening. I am just recruiting at my Alma Mater which is a traditionally all-women’s college in the Northeast (USA) with a strong engineering background. And I can tell you, there is a lot of excitement about working in energy.  Making those reliability changes from the inside. 

I am inspired everyday by my mentors, with 30+ years of experience with GE and Baker Hughes, they are the ‘good data’ providers. They have been able to make sense of those thousands of numbers that I work with, we need experienced people to add the science, the engineering and the purpose behind the data.