Ten ways to sharpen your Soft Skills as a grad student

In part 3 of the “Research for All” initiative, we will cover how to sharpen your soft skills related to the day-to-day activity as a grad student, research scholar, or even a research aspirant!

Specifically, you will learn about:

  1. How to deliver technical talks and the importance of local reading groups
  2. Internal and external collaboration, common email mistakes
  3. How to prepare for a meeting with your research supervisor/guide
  4. Work productivity and managing your mental health as a grad student

Who is the article for, and how soft skills may benefit us?

Anyone interested to learn or to improve their research skills.

The advice is based on my collective experience in academia and presents what worked for me after years of trial-and-error.

So, you are free to agree or “agree to disagree” with them. I would be happy to hear your views and what worked for you!

The aim is to provide a better starting point, so you do not repeat the same mistakes I did.

Even if you can only take out a few hours a week, I sincerely believe that based on my years of mentoring experience, it will at least help you upskill yourself and also provide value.

One question that I always ask myself in such a scenario:

Why not? What if the outcome is better than we expect!

Your feedback is crucial to help me further improve this article, so don’t hold back and let me know about your views in the comments or by simply mailing me.

A short recap of other articles in the “Research for All” series

Part 1: Motivation, research from a self-improvement perspective, life-cycle of a research project, problem ideation, reading list (8 mins. reading time. It is a separate article, so please give it a read)

Part 2: Problem verification, baseline setup, novelty (will be out soon!)

Part 3: Soft skills, including delivering a technical presentation, collaboration, and work productivity (current article)

This is a part of the “Research for All” initiative, aimed at promoting research awareness and making machine learning research more accessible.

A related article that may be of interest: How I had secured a research associate position during my Ph.D. and moved to Germany?

Let us begin…

Read, read and keep learning …

Read journeys of other grad students

How to prepare and deliver technical talks

I have covered this topic in detail in my previous blog articles (how-tomy past presentations).

I have explained it in the first half of the following talk.

Detailed explanation on how to deliver technical presentations and the life-cycle of a research project

Organize local reading groups

  • Listen to talks to experts in your field
  • Could be all supervisor’s students or just students working in the same field like NLP or Vision or Graphs or Systems
  • Get away from only a paper-focused mindset (High input, high risk, Sparse output). You will not know the right way at the start of your research.
  • Give 50% time on the project, 50% on skills and big picture stuff

Start internal collaboration with lab members

  • Instead of being a single student author, work with your Ph.D. friends in a professional capacity.
  • If you are the sole research scholar (most common) in a project and do not have prior research experience, it is more difficult to make timely progress.
  • You may supervise additional students from B.Tech and M.Tech programs, who may help you to run part of your experiments. It becomes time-consuming as you need to train them separately, and they are busy with heavy coursework.
  • Ph.D. is quite an isolated path, but it need not be completely alone.

Prepare well for supervisor meeting

  • Be proactive, and set up a recurring meeting with your supervisor every two weeks.
  • Be well prepared; your supervisor’s time is valuable. Prepare specific questions
  • Present results (if any) and ask specific questions for feedback
  • Your supervisor meeting is not an exam or evaluation of whether you worked in the last two weeks, so please treat it as a feedback meeting
  • Prepare meeting minutes on a shared document. Template: Points discussed — a conclusion drawn — TODOs before the next meeting
  • After discussion, mark out a reasonable set of action points that you will work on next two weeks
  • Use common sense, and do not take things to heart if things do not go as planned. It is normal at the start as it takes some time to sync and happens with everyone
  • Develop a strong friend circle, you will need it

Prepare well for a result update meeting

This is a specific type of meeting with a supervisor or collaborators where you were supposed to do some experiments and get some results to verify the current hypothesis

  • Template: Start with a summary of the problem statement, the conclusions drawn over the last few meetings, and end with the agenda for today’s meeting

Instead of presenting figures or tables directly, you can adopt this strategy:

  • Given the current hypothesis, we aim to verify whether X is associated with Y
  • To test that, we perform the following experiments
  • For experiment 1, we learn the following (present table and figures)
  • Before moving on to the next result, pause and ask whether they have anything more to discuss on this point
  • Remember, this is not a performance meeting but a feedback meeting. Design everything so that it is easier for them to give feedback

Setting work expectations as a grad student

  • Ph.D. has a high learning curve, so do not try to sprint it
  • Understand what stage of your Ph.D. you are in and what is expected of you
  • 1st two years (best case scenario) — complete relevant coursework, literature survey, problem formulation (ordered); submit your first paper around 1.5 years. The remaining years — you can figure it out on your own
  • Ask stupid questions. If a new topic or task is discussed that everyone agrees on, but you have no idea about, please mention that at that point
  • If needed, elaborate that you may require some time to understand it yourself or ask to be connected with someone who knows that
  • Asking for help at the start of a project does not make you look stupid
  • Plan ahead. Usually, when a paper deadline approaches, be prepared to give extra hours in the upcoming 4 to 5 weeks to submit.
  • I will also add that if you can go the extra mile or work a few hours, please feel free to do that. But you should also know when to take breaks. This is part of the learning curve.
Photo by Elisa Ventur on Unsplash

Stress Management

  • It is a tough topic and not easy. Learning to deal with it myself, so no advice on that 🙂
  • What worked for me till now
  • Follow Twitter handles that talk about common issues faced by Ph.D. students — #AcademicTwitter, @PhDVoice, @jenheemstra.
  • Before you start something, plan your goals well
  • Have long-term goal
  • Have short-term action items
  • May use the SMART framework while drafting your proposal or problem statement (Specific, Measurable, Achievable, Relevant, and Timely)
  • End goal: Clarity on the problem statement, input, and expected output
  • It is OK to say ‘No,’ but you give a reason and action point of what you will do instead
  • Setting reasonable expectations. This is much tougher than it sounds, and I am still learning about it.

Work Productivity

  • Taking Notes/ Journaling — day-to-day results and observations
  • I preferred taking notes in exercise books
  • Online version — Google Docs for meeting minutes, Collaboration progress
  • Detailed — Evernote, Notion

I recently came across a LinkedIn post and completely agreed with it.

You can’t manage what you don’t measure. And without measurement, real change is hard.

LinkedIn post

Photo by Solen Feyissa on Unsplash

Common email mistakes

Quoted from a Twitter thread, check it out for more details

  • Not specifying what you need
  • Not providing adequate context
  • Not being concise
  • Not following up

Pre-doctoral research fellowships

This is a good alternative if you are not sure whether you want to commit to a Ph.D. ( a long, arduous, and decisive journey)

Programming basics for ML research

Free courses are available on Kaggle called “Python” and “Intro to Machine Learning.”

Courses available at https://www.kaggle.com/learn

I have written an article on how to start using Python for Natural Language Processing, which may be of help to you.

Python data structure basics, data handling using pandas, and NLP tasks using NLTK Python package


This brings us to the end of the 3-part blog series “How to Keep Calm and Continue Research: A Self-improvement Guide” under the “Research for All” initiative.

You can find some of my old articles related to this topic under the “Research” menu on my website.

I hope you have enjoyed it.

Please let me know how you found that and what parts you agree or disagree with.

I want to conclude by saying that this series aims to bring awareness to known problems and solutions during your research journey.

As each of us and each situation is unique, I hope this awareness will make you feel less alone, more confident and calm, and last but not least, retain your love for research by losing your way less often.

One more point, you also need to learn when and whom to ask for help.

Image src: https://cs.stanford.edu/degrees/phd/PhD_Orientation_Slide-2021.pdf

💚30+ free articles already available at datanalytics101.com

💚 Your feedback is critical to improving the content, so please feel free to share your take on this topic

💚Follow me on Twitter @roysoumya1 for getting updates on “AI in Healthcare.”

💚I plan to write one post a month on Medium. To get updates directly to your email, please subscribe at https://medium.com/subscribe/@soumyadeeproy

What is your take on this topic?