Masters journey with upGrad, IIITB and LJMU

Photo by Markus Winkler on Unsplash

Introduction

In today’s technology industry it is of utmost importance to continuously up skill and be a life long learner. The recent pandemic has further elevated this need to enhance one’s skills and learn new ones. It has also given rise to an Online Learning format which was not very vividly used prior.

upGrad is one of the organizations empowering working professionals to gain some valuable and industry useful skills in multiple domains.

I have recently completed my Masters in Data Science from Liverpool John Moores University in association with upGrad and IIITB. Masters definitely increases credibility, opens up more avenues for jobs. Building a ML model base up also helps understand multiple aspects of Machine Learning.

This article will mostly be about my journey of Masters along with some tips for learners looking forward to pursuing this course.

The course is divided into 2 parts. 12months of Executive PG / PGDDS from IIIT-Bangalore followed by 7–8 months of Research in collaboration with LJMU.

IIITB

Masters starts with International Institute of Information Technology Bangalore PG Program. The initial 5–6 months are same for all the learners wherein the basics are covered , i.e. Python, SQL, Exploratory Data Analysis, Hypothesis testing and few assignments based on all of these topics.

Followed by this the next ~6months are specific to the specialization chosen by the learner and the recent syllabus offers 6 specializations which can be chosen by the learner depending on the career aspirations, interests.

I had chosen Natural Language Processing as my specialization. My interest in NLP was one of the primary reasons of making this choice which also helped in my Masters Research.

The pedagogy and the course delivery is in a blended form. There are weekly deadlines and modules which need to be completed. The learner is also expected to finish the set of pre-recorded lectures from University and Adjunct Faculty. There are live sessions organized very frequently on the corresponding topic which the student studies/is studying on the same weekend.

The assignments, case studies, viva, in-module questions, exams together decided the final grades and hence it is mandatory to complete all of them in the stipulated time.

Tips:

1. Try to get in depth of the algorithms, concepts being learnt (from resources other than upGrad modules), do more hand-on practice with different datasets. Kaggle is a great place to find datasets to experiment.
2. Be active on the discussion forums as it will help in resolving your doubts as well as learn something extra .
3. Plan to dedicate sometime daily (rather than saving everything for weekends/holidays)

LJMU

After 12months with IIIT-B a new journey starts with Liverpool John Moores University. First 1.5–2months are recorded lectures and weekly sessions by LJMU faculties which help in understanding various aspects of research , its types and how to do research.

The learner has to select 3–5 topics and submit them along with the datasets and brief write-up. LJMU will give a go-ahead for 1 or more topics based on the research potential. Next step involves reading multiple journal papers, articles, research material, etc. to better understand the topic, various methods which has been used by previous researchers and benchmark results.

Further the researcher is tasked with writing code to perform EDA, hypothesis formulation, modelling, algorithm tuning on the dataset (used in the research) to build a strong ML model. Based on the literature review and ML modelling the researcher has to write a Thesis document which is evaluated by LJMU. Supervisor is the most important person in the Masters as he/she will help you in planning, understanding, identifying important attributes and achieving the research objective.

Thesis might look like a daunting task at the beginning, but once complete it definitely gives a sense of achievement and the learning is extremely valuable as it builds the capability of writing a scientific document.

Few tips from my personal journey:

Research topic should be selected taking into consideration the interest , available time, resources and depth of understanding in the domain.

1. Following a schedule and incrementally writing the thesis documents (80–100 words) per day helps in avoiding last minute rush and ensures a good thesis.
2. Regularly update the supervisor on your progress and take feedback
3. Make a spreadsheet to capture the various hyper-parameters used and their performance (accuracy, precision , recall, F1 score, etc)
4. Keep track of the library version numbers (sklearn, tensorflow, etc)
5. Try to complete the first run of code (with decent results) as early as possible so that you have time to tune the hyper parameters and also enhance the scope of your problem statement.

Research Topic Selection Tips

The most important step in Masters journey is selecting a research topic .

Few pointers which have helped me in selecting my research topic :

1. Identify area/technology of interest — like NLP, CV, Deep Learning, Reinforcement Learning, Clustering, Classification, Regression, Data Engineering, Data visualization, etc
2. Identify domains (optional) — Retail, BFSI, Telecommunication, E-commerce, Manufacturing, etc. Domain understanding helps in understanding the topic, pain points, industrial application and usability of the research.
3. Self assess the skills — if you already are familiar and good with anyone of the areas and domains(mentioned in 1 and 2) then have at least one topic from that particular domain/area.
4. Navigate to the following websites to search for relevant literature (papers, journals, white papers, etc.):

Dataset Search (google.com) — Excellent resource containing the dataset location, papers citing the data, authors, last updated dates, etc

arXiv.org e-Print archive — Millions of papers, articles, etc

The latest in Machine Learning | Papers With Code

Top AI and ML blogs to follow in 2020 | by Suraj Bhala | Medium — An article written by me which has the list of few most important blogs (can be used to understand the current industry research direction, topics, etc)

Kaggle: Your Machine Learning and Data Science Community — Probably the biggest repository of publicly available datasets

Career Support

upGrad provides following services as part of career support:

1. A dedicated mentor to help you with planning your career objectives.
2. Access to upGrad Career Portal having multiple jobs which can be applied using 1 click
3. Mock interviews with industry experts to help with self evaluation and better prepare for actual interviews
4. Sessions from industry experts ranging from hard skills to soft skills
5. Professional resume building service

FAQ

1. Is it guaranteed to get a job from upGrad career support ? — upGrad does try its best, but the learner also should try other resources (like Naukri.com, LinkedIn, Hirist, referrals, etc. to secure a job, specially if he/she is aiming at a domain switch like Software Engineer to Data Scientist.)
2. Time to be dedicated per day/week — From my personal experience for data science / ML AI Masters a minimum of 12 hours per week is required.

If you are planning to enroll and to save 10,000–30,000 on the total fees or have any other questions please feel free to reach out to me on LinkedIn. (https://www.linkedin.com/in/surajbhala/)