Who can be AI Engineer
Recently someone asked me who is an AI Engineer and what to do to become one.
I think the job description cannot be defined. It is just so vast field that, you have to narrow down the scope to something like DL / ML Engineer or CV Engineer, etc.
But then I thought about myself. Can I call myself an AI engineer? If you have worked on AI-related technology and have engineered a project driven by ML or AI then you start considering yourself as a candidate.
Mindsets
I think there are two types of thinking of you can call mindsets.
- Engineering mindset or Industrial thinking
- Research mindset or Academic thinking
Industrial / Engineering mindset.
- Production oriented
- Product-oriented
- Can be associated with ROI (return of investment)
- Using already available resources, modules efficiently
- Optimization in terms of productivity and efficiency
- Fast-paced (can be short-termed)
Research / Academic mindset
- Research and basic principles oriented
- Highly Innovative and out of the box thinking
- New approach is generally more important than optimization
- Timed and persistent effort
So a researcher’s vs an engineer’s point of view over the same object differs significantly.
Take a simple example of a heart rate monitor or a traditional stethoscope. Engineering approach would be to optimize it mechanically, to maximize the sound captured.
A disruption in the traditional way could be brought by research by using captured images with a particular spectrum to analyze the frequency of intensity variation.
You still require an innovative engineering approach to design this new device and optimize it.
Research and Engineering always go hand in hand in all disciplines of science.
So coming back to AI engineer, it’s almost the same concept. One possible addition is that AI engineering or research is not limited to the specific discipline of science.
Anyone from any discipline of science can become an AI engineer.
Bridging the Gap
Here comes my favorite guy.
The AI / ML engineer
This guy talks with researchers and combines the engineering knowledge and connects research to production.
This guy may also work as a prototype engineer.
- No need to worry about scaling but keep it in mind while developing
- Making fast engineering prototypes of research ideas
- Agility to understand the research topic as well as engineering principles
Mechanical engineer to become AI engineer
When I started the Study Groups with MLTokyo, I got a wide range of participants for the machine learning study group. There were electronic engineers, mechanical, designers, finance guys, real estate people.
We always asked the first question :
- Why do you want to study machine learning?
And we got many types of responses
- To get more salary
- Catch-up latest technology
- Curiosity
- No reason
But there were some people who could connect and find a reason to learn AI.
This mechanical engineer guy was working in a pipe company. They installed water pipes at different locations.
In the case of pipe failure, the repairmen had these tasks:
- Physically to the place
- Capture and send images of the failed pipe back to headquarters
- Mechanical engineers would then analyze it in their 3d models for reason and solution
- Repairman goes back to the place and fixes it
The reason why this guy wanted to study AI especially Computer Vision was to make the life of repairmen easy.
An app with computer vision feature and ability to quickly predict the possible reason and at the same time suggest solutions based on data.
That would break the traditional industrial way of repairing pipes.
AI for everyone
Even if you don’t have much knowledge of AI, you can still learn it because it’s just investing your time and efforts.
Brainstorming to use AI in your field of engineering can guide you in this enormous sea of competitive AI world.
