Khushbu Saxena
Software Engineer
Squirro
What is Your Role and how did you end up in this field?
I am a software engineer at Squirro with a background in Artificial Intelligence specialized in text data. I became interested in the field of Natural Language Processing as it astonished me that a machine can mimic the magnificent ability of humans to learn from their experiences. Additionally, I am passionate about generating insightful solutions by studying the contextual difference in industrial data using Machine Learning.
I pursued my master’s from the Technical University of Munich focusing on Machine learning for text data. During this, I developed a solution for high CO2 emissions while training s.o.t.a transformer-based language models. We published TopicBERT for Energy Efficient Document Classification in EMNLP2020 (Findings). I also worked on the challenging task of fake news detection. I scored 2nd position out of 24 teams in the shared task Fine Grained Propaganda Detection in the NLP4IF workshop, EMNLP (2019), and have published a paper Neural Architectures for Fine-Grained Propaganda Detection in News. These experiences helped me to develop an in-depth understanding of the field.
How Would You Define Engineering in Your Own Words and What Inspires You about It?
Being an engineer for me is about being able to solve problems and being familiarized with minute details during the process whilst being able to look at a bigger picture.
What was the most challenging aspect of choosing this area of study?
The novelty of the field poses several challenges. As every field evolves, there are certain tried and tested best practices followed by engineers. The dynamic nature of AI is exciting but comes with a cost. The contextual variation in the data adds up to the difficulty. The rapid advancement in the field requires constant adaptation to the new higher-performing models and real-world industrial data is even more difficult to crack.
Moreover, being a woman in AI, one of the most challenging aspects was not having any role models in the field. While planning for the next level, I had no relatable examples to look up to.
What kind of impact do you make in the engineering environment at Squirro?
As a software engineer with expertise in AI, I bridge the gap between high-performing Natural Language Processing and highly reliable Software Development.Moreover, I wish to bring a diverse perspective to the engineering environment at Squirro. I believe that a diverse team is well equipped to identify and remove AI biases as they interpret data. According to Deloitte’s report, in order to build an effective AI system including defining a problem for AI to solve, designing a solution, selecting and preparing the data inputs, and constructing and training the algorithms, an AI team should be as diverse as the populations that its AI will impact.
Would you personally agree that attitudes and practices pertaining to employing female engineers are changing?
Yes, the change is definitely visible. Tech Companies are putting efforts not only into employing more female engineers but also into encouraging women to take up engineering and AI as a career. When I decided to pursue a career in STEM, I believed that me being a female has nothing to do with me being in tech. After all, one should choose a career on the basis of their passion, aptitude for the field as well as whether they enjoy it or not. However, having spent a few years in STEM, I have realized the importance of inclusion. I want to address why a company's attitude towards gender diversity matters. During the initial years of my career in STEM, I did face the lack of inclusion and had to constantly prove myself while interacting in a male-dominated group. Here at Squirro, I realized how inclusion helps me grow as a professional. It is important that my energy and attention are spent on my work rather than fitting in. I love working in team settings and my motivation is high when I work with people. In such scenarios, it is crucial that I don’t feel like a separate entity.
Given that there is a skills gap that does need to be addressed, what piece of advice would you give to all tech enthusiasts that are looking to start in the field?
While pursuing AI as an engineer, in-depth knowledge is the key requirement. AI is often considered a “black box” and it does not work all the time. To know why it doesn't, one should grasp the fundamentals very well. There are various courses available online which can be leveraged, such as the selection on the Squirro Academy. Additionally, there is a gap between research and what can be applied to real-world industrial data. Try to get as much experience with the AI industry that deals with real-world data.
