Naveed Ahmed Janvekar is a Senior Data Scientist working at Amazon in the United States. He works on solving fraud and abuse problems on the platform that impacts millions of customers of Amazon in the US and other parts of the world using Machine Learning and deep learning. He has 7+ years of expertise in the Machine Learning space which includes classification algorithms, clustering algorithms, graph modeling, BERT to name a few. He is using Machine Learning and Deep Learning to solve multi-faceted problems. He has a Master’s degree in Information Science from The University of Texas at Dallas where he graduated top of his class and was awarded as a scholar of high distinction and inducted in the prestigious International Honor Society Beta Gamma Sigma. He has a Bachelor’s of Engineering in Electronics and Communications from India. He has worked with other influential firms such as Fidelity Investments and KPMG. In his current role, he is researching identifying novel fraud and abuse vectors on ECommerce platforms and using Active Learning to improve Machine Learning model performance.
Editor’s note: The opinions expressed in this article are solely those of Mr. Naveed Ahmed Janvekar and do not express the views or opinions of his employer.
Q1: Tell us about your journey in AI and data science so far. What factors influenced your decision to pursue a master’s degree and a career in the field of AI?
Naveed: I am currently working as a Senior Data Scientist at Amazon, working on improving customers’ shopping experience by the detection and prevention of abusive or policy-violating entities within the platform. My journey in AI/ML started slightly before I enrolled in the Master’s program at UT Dallas. While I was working with Fidelity Investments in India, I was inspired by a couple of analysts who were making use of data to make impactful business decisions. This experience, along with my ambition for pursuing higher education, led me to study information science with a specialization in machine learning. After graduating from UT Dallas, I worked with KPMG as a Business Intelligence Developer working on building reporting applications. In 2017, I joined Amazon as a Business Analyst and worked my way upwards to become a Senior Data Scientist.
Q2: Tell us about your current role?
Naveed: My current role as a Senior Data Scientist involves building strategic Data Science roadmap/projects to continually improve customers’ shopping experience. On a day-to-day basis, I engage with various business stakeholders on various business problems and peer scientists to discuss the latest ML methodologies. Model building, experimentation, data extraction, and coding are pretty much part of the daily routine. Innovating on behalf of the customers is on the daily.
Q3: What are some of the biggest challenges as a Data Scientist?
Naveed: I believe one significant challenge is to get the right kind of data for exploration, model training, and/or insight generation. Many times, the data that is available might not be structured or even available in relational databases. There could be data quality issues, missing data, and features needed for model training might not be readily available. In addition, being able to engineer these features can be pretty time-consuming and complex.
Another challenge with respect to supervised machine learning models is the lack of availability of high-quality training datasets. By high-quality, I mean aspects such as enough volume of labels, data quality, and class balance to name a few. Sometimes building a narrative and effective storytelling of any analysis or data science solution to stakeholders can be challenging based on the complexity of insights. This is something one gets better at with time and constant engagements with business partners.
Q4: What is your opinion of machine learning in the field of fraud and abuse prevention?
Naveed: Machine learning helps in the scalability and accuracy of fraud and abuse detection in a cost-effective manner. For example: if one were to manually evaluate every transaction or entity as fraud or not then there is a pretty good chance of catching all bad transactions or entities. But in today’s world the scale of transactions and interactions is huge, billions of transactions, millions of entities make it humanly not cost-effective and possible to evaluate all of them manually. By using machine learning, one can automate fraud detection as much as possible with high predictive power and cheaper costs.
Q5: What would your advice be to budding machine learning and data science candidates?
Naveed: Data Science is considered as a generalist role by many. Hence while having data science breadth knowledge is important, my advice would be to also to focus on data science depth knowledge such as mathematical details behind algorithms. This will give you an edge over others in the field. Also, being good at storytelling, communication insights to business stakeholders, and building a narrative around your solutions. Participate in as many data science competitions as possible, participate in publishing research papers and get mentors early on in your career.
Q6: Can you name some books, courses, or other resources that have influenced your thoughts the most?
Naveed: My biggest influencer has been Andrew Ng, an American Computer Scientist. He has courses on Coursera which are pretty good. Hands-On Machine Learning with Scikit-Learn by O’Reilly Media is a good book as well. For programming skill improvement my go-to is leetcode for both python and SQL. The Data Incubator is good for fellowships and certifications.
Q7: What are your views about Marktechpost.com?
Naveed: I came across Marktechpost in 2021 and have been following it ever since. I really like the kind of articles and content that is being put out by this publication. Most of the latest ML innovations are covered and very nicely summarized for quick reads. This clubbed with the courses made available by Marktechpost makes it a great place for AI professionals to regularly engage with the platform.