In our interview series with the new generation, this time we interviewed Faisal Ahamed Khan. He is currently working as Head of R&D at Giga Tech Limited. Read his interview below:
First, we would like to know about you
I am an enthusiast of Natural Language Processing who is interested in solving real-life problems using Artificial Intelligence and Machine Learning. I am passionate about learning and adopting new technologies. I believe that working collaboratively is more effective than working alone when it comes to achieving greater things.
What is your research topic?
During my short research career, I have worked on several core components of Natural Language Processing. In addition, I have gained experience in areas of linguistics, especially shallow and dependency parsing, as well as contextual part-of-speech tagging. I also have extensive experience in data annotation pipelines.
Currently, I am leading a small team researching an end-to-end ecosystem for Natural Language Processing. This includes creating human-annotated datasets, developing language models, fine-tuning models for specific tasks, and ultimately building applications for end users. Among these tasks, one of our key responsibilities is benchmarking and reporting our results.
Language models generally serve as the foundation for other models and support them in solving complex and specific tasks. These models learn patterns within data, enabling them to provide generalized outputs. Their main task is to establish relationships between every word in a sentence, which then helps in producing meaningful results according to context.
How is your research benefiting us or will benefit us in the future?
We are trying to create and benchmark high-quality datasets for several core components of Bengali linguistics, which will serve as valuable materials for further important research. In our last paper published at KDD, the sentiment dataset was made openly available for research. The paper also presented analysis on the state-of-the-art accuracy of sentiment analysis models. When working with any model, published datasets play a crucial role in model evaluation. Moreover, a work regarding the development of rules for Bengali lemmatizer was published in EMNLP 2023 Findings. Currently, a research paper on named entity recognition is ongoing.
Do you have any messages or advice for young students in Bangladesh who want to pursue science?
Students in our country are extremely talented, but there is a lack of proper nurturing and coordination for their talent. My advice to them is to determine their career direction during their honors or undergraduate years. Instead of thinking about what to do after graduation, they should make plans at least a year before leaving university about what they want to become in the future and prepare accordingly.
Students should regularly read recently published research papers on their field of work. Alongside this, they should have an understanding of what resources are available on the subjects they are studying. There is no alternative to regularly reading research papers to prepare oneself for research. If students can review and present different research papers in a group each week, it will be even better.
Finally, what I want to say is that with belief and hard work, anything can be achieved.
Your email address: [email protected]
Your LinkedIn site address: www.linkedin.com/in/faisalahamedkhan
Your website, research links: https://scholar.google.com/citations?hl=en&user=jAk1N_YAAAAJ
On behalf of Scientist.org, we wish Faisal Ahamed Khan continued success. He is an exceptional role model for our young researchers.
Leave a comment