Abhishek Aditya Kashyap is a Product Manager at Blinkit with four years of experience in the Data Science domain. Abhishek graduated with a Bachelor’s in Chemical Engineering, where he earned himself the nickname, A2K, paying homage to every molecular nomenclature in chemistry. In this article, A2K discusses his story of how he switched paths from solely the data science domain to a more product-focused approach with a specialisation in data. A2K currently leads data-driven personalisation on the app.
Month 0: The Background
The idea of building and managing a product used by millions of people on a regular basis is super exciting. But, at the same time, I did not want to walk away from my four years of experience as a Data Scientist. I stood at a crossroads. In the words of my favourite poet, Robert Frost (slightly modified for context 😛) —
Two roads diverged in a yellow wood,
And sorry I could not travel both
And be one traveler, long I stood
And looked down one as far as I could
Innumerable questions of What, When and Why
I found my answer in the blink of an eye
The interviews were not only about gauging my fit in the company and the role but also about my expectations and aspirations. I can’t thank Adit, Nikhil and Richa enough for believing in me and bringing me onboard.
I joined the Personalisation team at its nascent stage, which offered me exactly what I was looking for — an experience of building from scratch as a Product Manager (PM) and an array of problems to tackle through Data Science (DS) and Machine Learning (ML).
Month 1: First Impressions
To begin with, the induction sessions touched upon every facet of the business and introduced all the internal tools available. The highlight of the onboarding process was a visit to a dark store to get a feel of the field job and unravel the magic of delivery-within-minutes.
My team is everything one could ask for — grounded in knowledge, experienced with a diverse set of skills and above all, fun to work with. Each and every one of them brings their own unique perspective to the problem at hand which builds a solid foundation for our solutions. It is a great learning experience to be a part of a team that is curious enough to identify new problems and work on them selflessly.
I was in awe of the fact that there are no hard boundaries across teams and everyone was more than welcome to contribute wherever they wanted to. The slack channels for feedback and the attention to each suggestion, made everyone feel heard. Bi-weekly town halls (that’s once in two weeks) and in-person visits by Albi (our founder), instilled a sense of ownership and belongingness.
Month 2: Data & ML
It was time to make myself familiar with the data. Besides doing exploratory data analysis to understand user behavior and trends, I started working on a recommendation engine. The focus was not only to build a personalised home feed for the user but also to enable discovery of new items and categories we offer. Iterations over internal brainstorming and business context from the leadership helped us finalise the first version of the model which was ready for production.
With all the relevant metrics in place and the deployment sorted out, we rolled out an A/B experiment against the popularity-based algorithm. Initial results showed a positive impact on category discovery and user adoption of new items. It gave us the confidence to roll it out to a larger audience and keep improving the model based on user reactions.
During this phase, I was fortunate enough to be involved in cross-team discussions as well, where I extended hands-on contribution. Being well connected across teams helped in learning from their experiences and transferring their knowledge onto our solutions.
Month 3: The PM Role
As someone who has walked in my teammates’ shoes, empathy came naturally to me. I understood their language which made the brainstorming sessions very interactive and successful. Guidance from within the team and beyond, helped me make up for my lack of PM experience and pushed me to rise up to the challenge.
Being a DS PM, it is my responsibility to take in numerous feedback pouring in from users and stakeholders and distill out the improvements required in the current algorithm. It also falls upon me to bring explainability to the model’s output that users see on the app.
The ability to explain complex technical concepts and convey them in layman’s terms to the business and the users is a hard skill to master. But, the end result is worth the effort because it allows you to gain everyone’s trust. Nobody likes black-boxes to dictate the state of a product. As soon as we attach ML black boxes with labels like “Because you bought …” or “People similar to you liked …” or “Because you searched …”, it becomes easier to understand and accept. Much better (and less painful!) than explaining user collaborative filtering or graph link prediction or contextual bandits.
This short yet impactful 3-month journey whooshed past. I feel privileged to be on this voyage with some of the brightest minds I have come across. I have learnt a lot in this short span and had an equal share of fun. The culture speaks for itself when you witness people leading from the front, irrespective of their years of experience.
I am looking forward to the exciting stint that lies ahead for me where Blinkit becomes a verb — Let’s Blinkit!
A2K is a PM at Blinkit. You can follow him on Linkedin.
Thanks for reading this blog. Say hello on Twitter or follow us on LinkedIn.