Hi there!
As title suggests, I am on the beginning of my career as an ML engineer (hopefully) , and I feel in need of some guidance. Firstly to give a bit of context who I am and where I come from… (please bare with me :))
I am an experienced Control System engineer with many years of experience working in the field of system modeling and controls.
Past year and a half I have dedicated a big portion of my time to finish several specializations on Coursera, mostly the classic from Andrew Ng, and few additional in CNNs and RNNs and working with time series. I also plan to finish one on databases and one on MLOps.
I am familiar with C, and Matlab, but past few years I have also worked intensely on learning Python and Keras/TensorFlow and since I have experience with programming it came to me with a lot of ease, to switch to using Python. Besides that I am no stranger to calculus, linear algebra, basic statistics etc, hence I have no issue understanding intuitively most ML concept as well as reading research papers.
I have toyed around with data series prediction / system identification applying RNN to model physical systems black-box style, used RNNs and CNNs for sound detection and categorization (distinguishing drums, transcribing melodies into notes etc). I have a few projects on GitHub and google colab that show I know how to implement things.
I did a few Kaggle entry level competitions to work with XGBoost, polynomial regression etc. Now I focused a lot on reinforcement learning try to grasp basic concept such as DeepQ, Policy gradient/PPO, actor/critic methods and I have implemented most of these on my own on basic examples from Gymnasium.
I wish to start using Isaac Lab after to tackle some harder RL problems.
My question in all of this is - where do I go from here?
Is there a need for me to narrow down and specialize in one domain rather then trying to be a Jack of all trades in ML and working on all these small projects of mine?
If so, what is currently looking like a subdomain of ML with most prospect of landing a job? (ofcourse my interests play a big role, but I find most ML interesting) Does RL seem to be a good field to specialize in atm? (it is also closest to my original profession, which is close to Robotics)
Should I team up with other people on Kaggle or here and start working on some common projects or am I ready to start applying for jobs and see what happens?
Does anyone even look if I wrote things on Medium or if I participate in Kaggle competitions?
I earn a decent salary as Control System engineer, so I guess I will have to downgrade to a near intern level in that sense, if starting ML, since I have no real working experience there (a little bit of application in my company).
I feel in need of some mentorship and also I would love to find people who would like to collaborate on projects. I guess having a concrete team projects behind me should help demonstrate I am not all that green Thanks!