When the necessity of technology utilization is addressed at the right pace, we experience exponential growth in our transformation. In the current trends, the quantum of research invested in Data Science is yielding humongous opportunities across multiple sectors. As a part of the Data Science interview session, we thank Dr. Sreekanth Mallikarjun from Reorg for sharing his technology dimensions and experience from his on-going successful journey.
In this conversation, Sree has shared some insights on below,
- The significance of “accuracy” with ML & NLP especially in the Finance Sector
- Result Focused attitude can turn out to be a successful strategy especially in the technology domain
Also, do catch up on interesting information on the recruitment lead for the Data Scientists.
Data Science: A Transformation
Jagan: How vital play does ML and NLP currently contributing to the Finance Sector where the ‘accuracy’ is the major parameter?
Sree: ML and NLP have long occupied a well-established place in the finance sector. Popular ML and NLP techniques used in the financial sector include time series forecasting and sentiment analysis. ML and NLP are necessary to mine value from the ever-expanding pool of unstructured text data.
While no data science model is 100% accurate, finance itself is a high risk, high reward, and often unpredictable business. The financial sector has benefited from incremental innovation and continuous advancements in data science methods.
For each new method we employ, we evaluate how much a new method improves accuracy compared to current approaches, and how much value that associated increase in accuracy generates.
Jagan: In your opinion, what have been the most relevant breakthroughs in data science impacting our world in the last 1-2 years?
Sree: I primarily work in the NLP space. BERT is a recent breakthrough that significantly impacted text-based data applications such as machine translation and feature learning.
Recruitment: Cheers Up
Jagan: What skills and attitudes do you look for when hiring data scientists?
Sree: An ideal candidate has strong analytical thinking abilities, a thorough understanding of ML and NLP methods, experience working with unstructured text data, and a graduate degree in a quantitative field.
Hard skills: To perform day to day tasks, one needs fluency with coding. Coding is performed mostly in R, while Python comes second, and it is also important to be familiar with SQL and NoSQL databases.
Soft skills: To succeed as a data scientist, one needs strong communication skills, creative problem-solving abilities, and business acumen. It is important to be able to break down a complex model into layman’s terms to engage stakeholders from different backgrounds.
Attitude: A data scientist should be self-motivated enough to fold up their sleeves and learn about the domain to identify underlying business value in raw data. Organization and attention to detail are also common traits among successful data scientists. Our team is highly diligent and collaborative and is receptive to feedback from any direction for the betterment of the product for our clients.
P.S. We are hiring now! Please reach out if you have these skills and attitude 🙂
Focus: A Successful Strategy
Jagan: One interesting tag in your profile is “results-focused strategy”. Could you share with us one success story of yours that uplifts this strategy?
It is important to have a results-focused strategy when faced with an opportunity to solve a problem that could produce a major impact on business. The complexity of the method or model is not important.
I was asked to help the commercial team better understand our customers and help them increase engagement with our product. The underlying data was scattered throughout various databases in different formats. I began by envisioning what an ideal solution would look like. This allowed me to conceptualize how to solve the problem and identify what results should be outputted by the model for the team to effectively act on. This was closely discussed with the technology and commercial teams to ensure that we all agreed. From there, I walked backward to achieve the required results. It began by harvesting data, some of which was readily available, but some of which needed extra resources to acquire. Using resources to obtain certain data required justifying why we needed the data and inferring how crucial it is for our results using deductive reasoning. Selecting a methodology and designing the model also depends on the size and nature of the data along with the expected format of results. Building a reliable model is a recursive process and obtaining continuous feedback from the business side is critical. In the end, the results helped the commercial team to predict customer churn early on and take precautionary steps to minimize it.
Motivation: A Step-up
Jagan: What were the impactful factors that drove you to pursue a career in Data Science (data & AI)?
Sree: The two factors that motivated me are:
- My interest in mechanical engineering and robotics. Constructing a data science model using software is similar to building a robot out of the hardware. It is exciting to follow through on an initial design and develop something that can be put into action.
- The many untapped opportunities within data science. This is one of the fastest-growing fields and it is a fantastic time to begin a career in an AI-related field.
Jagan: What is one book that you would recommend young data scientists to read?
Sree: I enjoyed reading Weapons of Math Destruction by Cathy O’Neil. The book gives a unique perspective on the ethics and limitations of data science while providing examples of useful applications.
Persona: A Well Defined
Jagan: If given a magical power, which global problem would you like to solve using AI/NLP?
Sree: World hunger
Jagan: Which technology use case has fascinated you the most in your childhood?