Let us get the introductions out of our way, shall we? I am Nirant, a final year Information Systems student from the desert college of BITS Pilani. I love Gulab Jamun, programming, data, new ventures and all related blabber.
I was interning with the Data Science team of a small startup Belong.co. This is that story, in a Q&A.
Wrong question. The correct question is Why Belong?
The answer is simple. We spend more waking time with our workplace teammates than with our spouses. We spend our youth, energy, love, sweat and blood building products with these people. Yet, we dont feel like webelong with them.
We intend to change this. As soon as possible.
Because I grew up looking up to the co founders. Vijay Sharma & Rishabh Kaul were people we would admire in silence. Stalk them on Linkedin with fake profiles, follow on Twitter and Yourstory. They are the Irrfan Khan and Nawazuddin Siddique of entrepreneurship. They took their own time.
And I love the idea of Belong. It is B2B business today and my mother does not get it. It is okay, it is my job to explain it till she gets it. She explained a lot of word problems to me till I got them.
Torture. Sometimes play with it. I dream about it.
It is Data. Dirty, mad, steel cold data. It is not big, it is small actually and I often fiddle too much and break something on playground server.
I work with the team called Discovery. We do what we claim to do that is help you discover right talent at the right time. I work more on the right time part of it.
I spent most of my time planning what to do and cleaning the data. I have some great mentors from both within the company and outside who are great help in this.
I split most of my screen time between a terminal and beautiful Jupyter notebook. Jupyter Notebooks allow us to interact with Python. It is my civilization of choice to torture data. The main combat units in this civilization are Pandas, Scikit Learn, and matplotlib. They are all from the Anaconda Python stack.
I sometimes prototype my models in R. It saves a lot of programming time. Though re-implementing everything in Python is not my favorite thing to do.
As a solution to this 2/3/4 language problem, I am considering using Julia as well. (Aside: JuliaCon India was awesome! If you were there, say Hi!)
Much less than what I wanted to. I have contributed to the design and deployment of a feature revamp which is important to us. It is a data driven product feature. It took me time to understand the context. Later, the architecture and what data models might work for this.
We shipped our first release last month. We have seen atleast 20% improvement in a key error metric. We also have better visibility of how the feature is doing. We can now move quick because we measure a lot of key metrics. Most of the metrics so far are around error measuring. We will be measuring operational, usage and some insightful vanity metrics as well. As a team, we firmly believe:
Merely measuring something has an uncanny tendency to improve it. -Paul Graham
I can write a pamphlet, wait, maybe a novella about this. But let me use a pun to describe this: I feel I belong with the team.
Most of the things I have bragged about above? Courtesy of ideas and wisdom from some of the best mentors I have. I have brilliant, helpful mentors from both within and outside the company.
My teammate, Sneha is the most kind, talented, smart programmer I’ve met. Did I mention she is a communist chauvinist Bong? She is fun to work with. Rohan Rao, Data Scientist and in-house mentor makes the most hilarious jokes at the most opportune moments. I think he should do standup comedy as a hobby. Unfortunately, he is India’s top Sudoku and Puzzle player. He represents India in World Championships et al. Nayan is the IIT KGP Informatics whiz from Bosch. He is the silent killer looking for the right time to throw the final punch. Vinay Pande is an experienced data scientist from Housing. I would write about Vikash, but that would mean this article becomes longer than you’d need. He has been like an elder brother to me.
In summary, none of what I did here is my work. It is the lowest common denominator of the gyaan that I get. I just test and experiment with the best of the ideas. Pick the one which works. Repeat.
Sorry, I can not answer this for you. I have not worked in any other company, team or startup before. Like first love, this will always remembered fondly by me. Even so, let me try to act grown up and try to write down my 3 favourite things about Belong.co — the team:
Everyone has access to all the code, all the required data, all the resources, all the help, all the company metrics. All the time.
The CEO talks about company revenues and targets openly and is not afraid to say, “It is my mistake, I will fix it”. His trust in kids like me is sometimes a burden in itself. Too big shoes to fill. #TrueStory
- Great Team Sync:
I am not saying we are perfect. But we are functional. There is great sync between all teams of the company even as we grew from 42 to ~80 Belongers. I am surprised sometimes by how easy the information flow is between tech and non-tech roles. The product development is much better because they make sure that customer feedback reaches us in an instant. Not 1 month later, not next week. Same day.
Several supportive, helpful, and meritorious people have some how assembled into Belong. I do not know how the founders did this, but they did. I have made some stupid but grave mistakes. Let us just say that I was lucky that I live to rectify those mistakes. I had timely constructive criticism and the time to improve on the same.
- (Bonus Point) Importance to Culture:
Most of the above is possible because the founders and leadership understand the importance of culture fit. We have never made a compromise on that. That spirit has percolated well and building a great team is an end itself, not the means to an end (aka profits).
Enough of Belong worship, see you later.