In hindsight, there would always be a bunch of things we wish we did / approached differently, wouldn't there be? To me, two things stand out particularly and if I could go back I'd tell myself these:
It is never too late; never too late to learn, never too late to switch, never too late to start on your own, never too late for anything. Most of what I am today and know of Data Science came from a decision I took when I was approaching 30 - an age considered too late by societal standards (especially Indian) to experiment. Though I'm glad I made that decision, I may have let gone of some others early on in my career because I thought it was too late.
Outward projection matters. Outward projection doesn't mean portraying oneself to be someone he/she is not. It's the attempt to accurately represent oneself for the strengths and ideas one possesses. The world and workplace today is highly competitive and it is more important than ever to know and believe your true potential and more importantly, project that potential appropriately.
There have been several short bursts of failure certainly, but one prolonged one that comes to mind. It is not a particular instance, or example, or project related or a promotion related failure, but a failure to carry my profession in the most effective way that I could. During a significant portion of my early career years, I was the type of guy who just accepted status-quo at work. I never really put in enough thought into what kind of work I liked or what kind of career I wanted to build for myself. I just merely accepted everything that came my way, because hey, I was doing quite well anyway and getting all the promotions. I did not challenge the directions in which projects went, the direction in which my profession went, or demand for better work or better pay. Mostly I was subconsciously just staying in the comfort zone of doing nothing drastic or taking decisions that felt uncomfortable but necessary. This went on for a few years until one day the harsh truth really hit me hard - that I had no real skillsets to talk about (except putting pretty powerpoint presentations perhaps). This day of awakening had a real profound impact on me. I couldn't deal with the fact that I had spent all those days and nights toiling away at work but at the end, I had nothing to be really proud of. That's when I decided to change the direction of my career. I did some really deep, hard thinking about the kind of professional I wanted to become and be known for, and the type of expertise I wanted to develop. Ever since, I've made conscious decisions on my career and work. I've spearheaded projects and challenged everything that could be challenged.
Most importantly, I hate status-quo now, which I think is a great attitude to have and I openly endorse it. This failure, and the subsequent realisation of it was truly essential for me to get a handle on how I wanted things to progress.
As with any career, ups and downs are natural. Perhaps not overwhelmed, but I have had my moments of feeling low, especially while growing and trying to contribute as a senior and later as a lead data scientist. Today, there has been profound improvement in the knowledge and understanding around what data science can bring to the table and how companies can transform themselves by being data-driven. Quite a few companies today are in a position to implement this knowledge and reap benefits from it. But this was not the case a couple of years ago - there was a lot of buzz for sure and the race to have the best data scientists was on, but good data architecture was seldom in place. Leadership found it hard to get the ball rolling on data science initiatives. In a debate between having a long term data science project kicked-off vs. some other sales or product sprint, the latter often took the cake. These were really frustrating times. Over the course of the next couple of years, I developed better leadership ability in terms of presenting more accurate value propositions from machine learning driven work, developed better prototypes to convince leaders, and in general, transform myself into an inspirational leader rather than be a data science project manager. Slowly but surely, Data Science work got better acceptance and we went on to create immense impact through quite a few projects.
I'd be building digital/data-driven products to uplift the not-so-fortunate sections of the society, do something with regard to spreading awareness about climate change, empower animal shelters to provide better care, nutrition, and a permanent home for animals. Technological innovation, internet, easily available data and scale of computing has made it possible for anyone sitting in any part of the world to create huge impact to how the world moves forward. Animals and lower rungs of the society barely receive any attention at all. Time to make a difference to those lives as well.