Some of my former PhD students started their new chapters as academics (assistant professors in a university). We had some personal conversations about things to pay attention to when starting a research lab in robotics. They encouraged me to write down some of the advice to reach a wider audience. Please note that what I have written below is more relevant to assistant professors in robotics.

Negotiate how you start: It is in the best interest of the institute that you succeed in research. Your research success feeds into your teaching quality in the long run. You cannot do research without good grants. Grant writing needs relatively uninterrupted time to do basic simulations and experiments. Very often one good grant proposal takes about 6 months to write even if you are experienced. This means, the first year of your assistant professorship should have more time to do research relatively free from admin and teaching duties. This should be supported by a reasonable amount of startup money and at least one PhD studentship to survive till you secure a grant. Negotiate these before you start.

 Things to keep in mind when hiring the first few PhD students: Your first few PhD students can make or break your career. Be extremely careful in hiring. Focus on what they have accomplished so far and their attitude towards research. Try to avoid those who need detailed instructions to take every step. They will drain a lot of your energy. Choose those candidates who have developed self-initiative, insight, and openness for a quality discussion. Otherwise, you won't enjoy doing research with them, and you will end up doing their PhDs. A good way to test this is to give them an assignment before the interview, such as a paper to read. Ask them to critic the paper and propose what they would do differently and any new questions they could propose to test. Those who have the right attitude and approach will be seen better this way.

Things to keep in mind when hiring postdocs in the early stages of the lab: A productive postdoc can hugely accelerate the progress of the lab. Manage expectations in your first postdoc. Be clear that the lab is still being developed, and the postdoc will have to contribute to overall management of resources in addition to their own research focus, or anything else you expect them to do. You have to find somebody with openness to explore your new directions. Therefore, be very clear about the research directions and check how open they are to be flexible. I would definitely arrange a telephone call with their former supervisors. You need to take what they say with a pinch of salt, but you need to have the instincts to be on the safe side.

Those who have developed predatory habits can end up demoralising junior researchers in your lab. For instance, some candidates can come from research cultures where a little contribution such as sharing a piece of equipment or giving a little advice is enough to warrant themselves to be a co-author of a paper that took several months of hard work by a junior researcher. Try to write down norms and expectations like I have done here for postdocs, and general supervisory approach written here. You can ask them to read such guidelines before the interview and see what they have to say about them during the interview. You need to have the instincts to spot the right answers.

Take management training: If your institute provides line management training courses, I highly recommend to take them. I benefited from courses that included frameworks to plan and conduct difficult conversations. Frameworks such as V.I.T.A.L and transformative communication has been very useful. In essence, avoid micromanagement and practice challenging your team to think for themselves.

Excellence in teaching: You may not realize that teaching is a great source of invigoration. Just the feeling that you pushed a bunch of young and aspiring students towards being able to solve technical problems that they couldn't handle before is a great feeling to end the day. It takes a lot of time to plan and execute a good taught course. You can take the first year to gradually build a good course. Focus on learning outcomes pitched at the level of students (i.e. 1st year or 4th year) and work backwards on the constructive alignment of course content and assessments. Design enough diverse tutorial exercises. Do not be discouraged even if you put a lot of effort but student feedback is not up to the level you expected. Reflect on the student feedback and think about revisions to content and methods of delivery. You won’t get everything fixed in one year. Usually it takes about 4 iterations (4 years) for a new course to become mature. Having got one course right, your second course will become successful faster. Remember, what matters most is your ambition to give the best learning experience to your students. Don’t confuse this with pleasing them. Your job is not to take a political approach to please them, but to challenge them enough and guide them to be successful in the long run.

Collaborations: You cannot do everything alone. You must build a reliable and productive network of collaborators. Here, be careful. You don’t need seniors just for the sake of having them in a grant or in a workshop you would organise. Carefully choose people who have complementary skills and time to spend on writing grants with you. Limit these collaborations to a manageable number in the first few years. Too many will thinly distribute your attention. I suggest you try about 2-3 good grant proposals in the first year hoping at least one would work out. But the first few grants you write should set the foundation for your lifelong scientific dreams. Therefore, you should take the driving seat in at least one, and be a co-investigator in others a senior person would write.

International leadership: Your new academic role is a great platform to contribute some visionary input to the community. A good way to start is to submit a couple of workshop proposals to the most impactful conferences in your area. In robotics we respect ICRA, IROS, and RSS. There can be other specialised conferences such as RoboSoft. It is useful to lead a workshop with a good team to explore ideas you are interested in. This sends the message to the community that you are passionate about pushing the boundaries of a particular theme. Usually, such visibility helps to elevate the credibility of the pathways to impact in your grant proposals. But please don't do this as a tick-box approach to pass tenure. If you prioritise workshops to align with your true passion, everything will start to work in your best interest.

Learn to say NO: You may get very convincing emails requesting your contributions to various good initiatives. You need to be on top of a sound priority list and know where to draw the line. Otherwise, very soon, you will feel exhausted and lose focus on things needing your full attention. Learn to write a polite response like “thank you very much for the interest. However, at the moment, I do not have the capacity to be part of this effort. However, I suggest trying A, B, C”. If you think you may consider contributing in future, mention roughly when they could remind you.

Avoid politics in science: You may see some people making fast progress by doing politics. Remember politics has a short life in academia. True science finally survives. Just focus on your scientific passion and be a good science communicator. You can supplement your papers with open code and data releases, Youtube videos of cool demonstrations, and public discussions etc. I recommend to spend around 5-10% of your time toning down your science to reach out to a large audience. This is a much better way to get your scientific outcomes spread across the world than doing strange things to boost citations.

Hope the above had some useful hints. Steer ahead and good luck with that tenure!