RTI International (https://www.rti.org/ https://www.rti.org/) does the type of work you're talking about. My understanding is they also pay better than what yo

RTI International (https://www.rti.org/ https://www.rti.org/) does the type of work you're talking about. My understanding is they also pay better than what you'd typically see in some policy institutes or postdoc jobs. They have Data Science jobs that don't require a PhD. Keep in mind with areas like public policy or other areas typically involved with government or NGOs, it's important to understand the subject area you are working in beyond pure quantitative analysis. Here's an example. I recently met somebody who is involved in a project that works with a government agency in Asia to build a model to predict the onset of a particular seasonal disease. The goal is to be able to predict where to allocate beds, doctors, medicine, etc. The statistical model they are building for this is hard to build well. It's this way because the data collection process is still being refined and there are politics and money involved with the spread and treatment of the disease. No magic algorithm is going to build a great model. They really need somebody who understands the data, the geopolitical climate, and the medical system to build a solid and useful model. You can train somebody in 4 weeks to blindly try algorithms and pick the model with the lowest error. But only somebody with expertise in at least one areas (e.g. medicine, government) can handle the other parts of this project. So they don't need a pure data scientist. They need a PhD or MD with expertise in other aspects of this particular problem to work with a statistician. It's hard to really get the traction you need in this type of problem without a PhD, MD, JD, etc. If nothing else, take a solid sequence of courses in statistics. If you don't mind, I'd also like to offer a counterpoint to your thinking about what makes a Data Science job meaningful. I used to work in Data Science consulting and worked with a lot of different companies. Some of these are perceived to be much more evil than good (e.g. Oil and Gas, beverages companies, etc). But you can do good work in these companies despite what they do in the larger scheme of things. For example, we helped an Oil Company build tooling for sensor monitoring so they could prevent chemical spills. In a small way, this helped the environment a little bit. For the beverage company, we helped their supply chain so they were driving fewer miles - a little less crud in the air and less fuel consumed. Not all of this was super-duper ML. Some of it was mundane BI. I'm aware some of this is rationalization. I've worked on "Personal Offers." This is portrayed as an extreme focus on customer satisfaction, but it's really just about getting people to buy stuff. I didn't enjoy it. Facebook has plenty of this kind of stuff, but they also have their data centers. Maybe a data scientist can help figure out how they can save some electricity and heating in those data centers, or keep servers running for longer so less hardware ends up in a landfill.