I joined FAYM Co in my third year, my first time shipping work that sat in production. The engineering team was small: Rajnesh, the CTO, Harshit on the frontend, and me. This is what stuck: the product, the code, the pace, and the parts you only feel after you have actually sat in an office.
Should I even join?
Before I said yes, I had the usual doubts. The setup was hybrid, three days in office and two from home, with about a two hour commute each way. I had wanted a startup early on, partly from creators and educators online (Harish, BlueLearn, Aman Dhattarwal, and others) who paint startups as a shortcut to exposure and real problems. The schedule and travel did not look great on paper.
I am glad I went. I already cared about frontend from my own digging, exploring, and breaking things until I understood them. That mattered from day one. I still remember opening the codebase and running the site, equal parts curiosity and mild panic.
What we were building
The idea was to list influencer referral products in one place so creators could share everything through a single flow instead of scattering links across apps and chats. When I joined, the app still had room to grow visually, which is normal at that stage. The experience was tuned for phones first, because the audience skewed heavily mobile.
First week and the screen that pulled me in
Early on I spent time turning specs into UI and wiring pieces together. One area of the frontend hooked me: the screen split into two sides. On the left you changed things like the referral link, title, and image. On the right you saw a live mobile preview. Debugging that flow and keeping both sides honest felt like real work, not a tutorial.
The interaction was deliberately fiddly. Fields looked like static preview text until you tapped or clicked, then they became real inputs. After save they went back to read only. We also kept rules like only one field editable at a time so the screen did not turn into a mess of focused boxes. ChatGPT was not yet the default pair programmer, so I worked the logic myself, which was oddly satisfying.
Side work: AWS, Streamlit, WhatsApp templates
I also picked up AWS adjacent work: read the Streamlit docs and built a small prototype to explore automating WhatsApp messages using approved templates, including listing a few OTP related cases, to help the team experiment. Glue work still has to be correct: APIs, templates, and guardrails matter.
Backend, analytics, and a CTO who listened
I read Express code, traced how data moved, and focused on the analytics schema. I wrote up what I found and walked it through with Rajnesh. The moment I still remember is when my notes lined up with something he was already thinking about. He came back with a revised schema that stored the same signal in less database space. For an intern, that is the good kind of intimidating: your homework mattered, and it tied to real cost.
Expectations, one bad day, and a useful lecture
Over time the team expected more, and for a while I kept up. Then one day everything felt chaotic and I could not untangle what was blocking me. The next day I got a direct, useful talk about how work actually happens: when to stop spinning, how to ask for help, and how to order tasks when everything feels urgent. That conversation mattered as much as any feature I shipped.
People: lunch, cricket, table tennis
It was not only tickets and pull requests. We had team lunches, a night cricket match on turf, and a table tennis table to decompress after work. Small things, but they matter when you are new. They turn the office into faces and jokes, not only a commute destination.
One late night, scraping, and side by side errors
I remember a task where I wrote a script that pulled data from a sensitive, NDA covered source, stored it in the database, and surfaced a side by side list that highlighted what failed. For that stretch we stayed late, and dinner was on the team. Rajnesh also nudged me toward breaking problems down, using markers, and writing pseudo code before code. That habit stuck.
Harshit was the person I leaned on when frontend got stuck. He explained clearly, brainstormed with me, and never made me feel slow for asking. I am grateful for that patience.
Calling people “bhaiya”
Both of them were mildly annoyed at first that I called them bhaiya. It was habit, not disrespect, and I found it hard to switch. After about a month Harshit was blunt: this is not home, bhaiya and didi do not apply here, use names. I started using names. It still feels a little awkward, but I can do it now.
Ending the stint and college overlap
About two months in, college ramped up and I could not continue full time. We tried one more week to see if a lighter schedule worked. It got too chaotic to balance with coursework. Office rhythm and campus rhythm do not overlay cleanly.
I still carry a college project from IT where the idea was to split an array of strings into two rows with minimal wasted space. I pulled in how I thought about layout from the frontend side. Classroom exercises and real software share the same habits: state the problem, handle edge cases, ship something you can explain.
The honest footnote: commute
What wore me down was long crowded bus rides at night and time lost on the road. Hybrid helped on paper, but heavy travel still drains you before you open the laptop.
Gratitude
I am grateful to FAYM for the learning window. Working next to Rajnesh and Harshit shaped how I think about shipping, asking for help, and staying calm under deadlines.
Thanks to Harshit for patience and clear explanations on the frontend, for brainstorming together, and for the standard he held without making help feel like a favor.
Thanks to Rajnesh for hints on tasks, for pushing pseudo code and markers before code, and for showing how to look at the whole project first, then break work into a sequence I could finish.
Thanks to Aman, the CEO, for building a place where that kind of mentorship could happen. I hope the company keeps climbing.
Those two months taught me more than I would have picked up alone. I will miss the lunches and the occasional cricket matches as much as the pull requests.