Unstructured data hidden in emails, blogs, social media, audios and videos gives us a truly comprehensive understanding of the bigger picture unlike structured data !
Niraj Kumar, our next pathbreaker, Research Manager at Fujitsu, works on R&D related to Industry Oriented AI, Core Deep Learning and Graph Neural Networks in areas that are directly or indirectly linked to human wellbeing.
Niraj talks to Shyam Krishnamurthy from The Interview Portal about his PhD on Intelligent Text Mining and his diverse range of work in the industry at the intersection of Human Language Technology, Deep Learning, Machine Learning, Social Media Mining and Artificial Intelligence.
For students, deciphering natural language is complex, because there are several challenges related to linguistics, syntax and semantics, which is what makes it a promising technology for the future !
Niraj, tell us what were your early years like?
I spent my childhood and did my basic education in a small town, Giridih (Jharkhand). I did my schooling in Saraswati Shishu Vidya Mandir. In the early years, I was too weak in Mathematics. But, the use of Vedic Mathematics and ancient mathematical tricks in education increased my interests in the Mathematics and Science subjects. I am thankful to my parents and teachers for the exposure they gave me during those days.
What did you do for graduation/post-graduation?
I did my graduation, and post-graduation from IGNOU (i.e., BCA, and MCA).
After schooling, I joined government colleges for higher education. Initially, I was not successful in the engineering entrance examinations, but due to my interests in Mathematics and Science, I opted for BCA and MCA courses from Open University – IGNOU. My parents fully supported me during my initial struggles.
After my MCA degree, I came to New Delhi for better career opportunities. I started teaching BCA, MCA, and BTech (Computer Science) students in different coaching institutes. I even got a few opportunities to work with IT startups.
At that time, I was not satisfied with my career progress, so I talked to my parents regarding an option for higher studies. At that time, they again encouraged me and gave their 100% support to my new journey. I had appeared in GATE and multiple Post-graduate level entrance examinations. Finally, I selected the “MS by Research” degree from IIIT-Hyderabad, which later progressed into a PhD in Computer Science.
The research and teaching environment of IIIT-Hyderabad totally reshaped my career. It made me an independent learner and researcher (i.e., made me ready for all fights in the future in R&D). So, I am highly thankful for my Faculty Advisors and Professors at IIIT-Hyderabad.
What were the influences that led you to such an offbeat, unconventional and unique career?
My parents and school teachers were my early influencers, who helped me overcome my fear of maths. They shaped me in such a way that maths and science became my strength and the most interesting subjects.
Finally, my faculty advisors at IIIT-Hyderabad, and professors, helped me in developing proper research thinking and made me an independent researcher.
My decision to go for higher studies has helped me a lot.
How did you plan the steps to get into the career you wanted? Or how did you make a transition to a new career?
Since my early days, I was thinking of making my career in the research field. The IGNOU BCA and MCA programs were good quality application-oriented programs. With these degrees, it was easy to crack the entry-level developer job applications. Even, I had got a few offers from IT startups. I had gained a good reputation in teaching BCA, MCA, and B Tech (Computer Science) students. But, I was not satisfied with that type of career. So, I decided to go for a higher degree program.
I had cleared the PG-level entrance examination and based on my performance in the exam, I got admission to MS-by-Research Program. Later, I decided to convert it into a PhD program. For that, I had gone through screening interviews and qualifier exams.
During my course work, I found NLP and Advanced Graph algorithms as interesting fields to study and research. Based on that, I decided to develop core NLP techniques for different text mining applications. The limited use of linguistic resources and social media-based graph algorithms was the beauty of all such algorithms. I had applied these core techniques to multiple text mining applications like: (a) Document Summarization, (b) Summarization Evaluation, (c) Document Clustering, (d) Keyphrase Extraction, and (e) Automatic Question Answering. My PhD thesis title was – “Towards Intelligent Text Mining: Under Limited Linguistic Resources”.
Achievements (during PhD program, this also proves the effectiveness of the work done during the PhD program).
- IN TOP SYSTEM @ TAC-2011: My system was in the top system for “Automatic Summarization Evaluation Task” at Text Analysis Conference (TAC 2011), organized by National Institute of Standards and Technology (NIST), Gaithersburg, Maryland, USA TAC-2011.
- IN TOP SYSTEM @ TAC-2010: My system was in the top system for “Automatic Summarization Evaluation Task” at Text Analysis Conference (TAC 2010), organized by National Institute of Standards and Technology (NIST), Gaithersburg, Maryland, USA TAC-2010.
- Other Recognition: Our Unsupervised Phrase Identification technique for Keyphrase Extraction, has been appreciated by the survey paper published in COLING-2010, Titled: “Conundrums in Unsupervised Keyphrase Extraction: Making Sense of the State-of-the-Art”.
As the nature of my thesis work was highly related to the industry requirements, I didn’t face any problem in getting the industry job.
I have around 9 years of industry R&D experience.
Tell us about your career path
I started my career as Research Intern at IBM IRL (Bangalore) where I developed a new system to answer – “Why based questions” with the help of Wikipedia dump.
After the completion of my PhD thesis work, I joined TCS Innovation Lab as a researcher.
Below are some details of my work at TCS.
- Automatic Plagiarism Detection system (highly effective in the case of paraphrasing and major word re-ordering). (winner of the best paper award 1st place @ CICLing-2014)
- Automatic text-quality grading system (Effectively grades the quality of E-mails, without relevant background model).
- Automatic event detection, alignment and prediction system (related to the economic events).
I also took up a Post-Doctoral position at University of California (Davis) where I worked on the problem of “Automatic identification of taxonomy of knowledge from software engineering documents”.
My next role was Senior Machine Learning Scientist at Phenom People, an AI-based SaaS platform with a purpose to help a billion people find the right job and discover their true potential. Here, I developed a novel weighted knowledge graph for Smart B-2-B-2-C hiring, for Contextual and Personalized Search.
I also worked on Candidate Social Graph for automatic resource arrangements and supporting business planners and Automatic Job-Highlight System.
In my next stint as Scientist at Conduent, an IT company. I I worked on problems in Conversational AI, NLP, Deep Learning, and Machine Learning. Some of the real world applications of my work were in Automated BOT-Humanization System, Deep Learning based Fully Automatic Crime Volume Prediction System, Automatic Hate Target Identification System, Fully Automated Cyber Hate Profiling System, Quantification of emotion and sentiment in calls and chats etc. Some of these were awarded US patents.
I then joined Samsung Electronics as Senior Research Staff and Manager. Here, my research work was in Conversational AI, NLP, Deep Learning, and Machine Learning. I worked on the complex cases of conversational AI systems, with Uni-lingual and code-mixed language support. We filed two patents on conversational AI and Smart systems. [Also recognized by the company as A-grade Patents]. We were also working on an emotional intelligence supported conversational AI system.
Iam currently working as Research Manager at Fujitsu India (FRIPL) where Iam working on R&D in the area of – Industry Oriented AI, Core Deep Learning and Graph Neural Networks
I always keep my eyes and ears open for key trends in research. I have always spent time learning new technologies and keeping myself up-to-date.
How did you get your first break?
I applied to TCS Innovation Labs for a research position and got selected after 5-6 rounds of interviews. It was an off-campus selection. It was a very good experience to work with people at the TCS innovation lab.
What were some of the challenges you faced? How did you address them?
If you want to be a good researcher and lead society, then you may not expect a lot of support. At the very early stage, you may not get sufficient materials and resources for your research. You may not get good tutorials to learn techniques. So, this is the type of real struggle that you will face. Only being a self-learner, having strong willpower, and having a clear vision can help you.
Where do you work now? What problems do you solve?
I am working for FUJITSU. Here, our goal is to come up with remarkable achievements in the area of AI and Machine Learning so that we can make revolutionary changes in almost all areas that directly or indirectly affect humanity. More details are available at ( https://www.fujitsu.com/global/about/research/ ).
What skills are needed in your role? How did you acquire the skills?
It depends on the nature of the job. Keeping up-to-date is the only way.
What’s a typical day like?
It is similar to any other working people. I usually start my routine daily activities, and after that, go to the office. During my office hours, I try to do my best to solve the research problems. Finally, after returning from the office, I spent time with my family. However, before going to bed, I keep 2-hours every day for updating my knowledge and learning new technologies. During the weekend, I prepare tutorial videos, where I mostly cover high-end technical topics. You can find these activities on my YouTube Channel
What is it you love about this job?
Response. (a) Continuous Learning, (b) Solving cutting edge research problems, and (c) proper recognition by corporate and research communities.
How does your work benefit society?
In the company R&D environment, each of the companies has its own philosophy to serve humanity. If you go through the corporate websites, then you can easily see this. However, these problems reach a researcher as complex or sometimes as completely new research problems. So, the role of a researcher is like a driving engine for the entire system.
Your advice to students based on your experience?
If you have real interests in a topic or area, develop your interest in that topic/subject.
My future plan is to learn a lot of new things and accelerate my contribution to game-changer research activities.