In high school I had gotten really into coding and had also seen a bunch of YouTube videos on quantum physics which I thought were super cool and unlike anything else I'd learned about. Once I heard about quantum computing this idea of using the crazy physics properties to accelerate computing really resonated with me. Definitely also a gut feel like yours at first Ari.

@Ari Noori For sure bro. I would say in my experience being impactful at a startup has boiled down to these 3 things; these are nontechnical skills because I feel like anyone with enough technical foundations that's passionate and has these qualities will be able to learn by themselves on the job. 1. Hustle. In a small startup you're talking to the CEO daily, they see who's clocking out at 5 and who's putting in the extra effort to hit deadlines no matter what. Someone who focuses on getting stuff done instead of # of hours worked, and that can work and learn independently without being a time burden on the CEO is super valuable. My boss has mentioned these qualities multiple times when comparing the few employees at our company and who he wants to keep long-term. 2. Agency. Don't feel like you have to ask for permission for every small thing you're doing, have a bias towards action while of course talking to your co-workers and asking for advice/approval when really needed. At a startup you'll provide much more value if you look for ways to improve your product/research that go beyond just what your boss told you to do; your boss has a lot of other stuff on their plate other than building the tech, and you likely know way more of the details about what you're building. 3. Humbleness. Especially if you're an undergrad with less experience, if you can financially you should be willing to take lower pay initially than a different job you can get, knowing that working in this industry will pay off long-term. Getting to work in a job that's conventionally for more experienced/older people will definitely pull you up in your career, and once you prove yourself you will get compensated fairly or now be able to apply for similar roles elsewhere and get the better compensation. People often say you should give FOMO to employers instead of acting desperate and that is completely true once you have experience, but when you're early in undergrad it makes sense to explain how you want to learn with a certain job and that you're willing to take less to start.

@Ari Noori I appreciate all your guidance when I was tryna figure this situation out bro 🙏🏼🙏🏼

@Ari Noori I think that useful quantum software projects would focus on software for quantum architecture or applied algorithms. In quantum architecture, one thing I'm familiar with is that there are tons of error mitigation (EM) methods (ZNE, twirling, DD, ex. in Qiskit) each with many options, and not always concrete recipes for which ones work best in which quantum algorithms / applications. So I think it could be valuable to pick a certain application of quantum computing and play around with the EM methods to learn about them and see which works best. In industry I've seen both these "user defined" EM methods being applied as well as companies trying to benchmark different commercial EM providers like QEDMA and Q-CTRL (probably hard to get your hands on those but developing skills with open source EM tools like Qiskit and Mitiq builds the same skills). For quantum applications, I'd pick an industry of interest like chemistry, material science, finance, or logistics. You can focus on either near-term algos that can solve problems in these fields such as VQAs (QAOA, VQE), or ones that will be usable later with fault-tolerant quantum computers like QPE. You can likely find a toy example on Pennylane / qiskit tutorials or someone's github. It would be cool to then find a harder problem in the industry that has not yet been solved with this algorithm in your field and play around with solving it; that really emulates work you may be doing in a company that's trying to do a proof-of-concept of these algorithms for their use-cases to show how they will extract value in the (hopefully near) future. Don't expect to get a quantum advantage yet haha, but rather showing the potential for quantum algorithms in that certain application.

@Devesh Vedantha A cool idea would be a ML model that gives a error mitigation strategy given a circuit as input. I was discussing this with a Prof at UW-Madison and this would help close gaps in extracting value from EM.