From Industry ML Engineer to Grad Student: What Changes

I got my acceptance letter from Northeastern a few weeks ago. It's sitting on my desk in Bangalore, next to a half-eaten packet of Parle-G and three notebooks full of GRE vocab I'll never use again. In about seven months, I'll be in Boston starting an MS in Computer Science.

Honestly, it still doesn't feel real.

The Mental Shift

For the past two years at Myelin, I've been shipping. Production models, CI/CD pipelines, customer demos, the whole thing. You build something, it goes live, people use it. The feedback loop is tight and addictive. Your pull request from Monday is improving someone's workflow by Friday.

Grad school is a completely different game. No PRDs. No sprint planning. No "can we ship this by Thursday?" The output isn't a deployed model, it's understanding. Papers instead of products. Assignments instead of production deploys.

The thing is, I've been feeling the gaps in my knowledge for a while now. At work, when a new architecture paper drops, I can implement it, sure. But understanding why certain design choices work, the actual mathematical intuition behind them? That's where I get shaky. I want to fix that.

What I Hope to Gain

Deeper theory. I want to actually understand optimization landscapes, not just call loss.backward(). I want to read a paper and follow every derivation, not just skip to the experiments section.

Research exposure. I've been a practitioner, an engineer. I want to know what it's like to push the boundary of what's known, not just apply what someone else figured out.

The network. Boston has MIT, Harvard, and a ridiculous concentration of AI talent within a few miles. The people I'll meet, classmates, professors, folks at meetups, that matters as much as the coursework.

What I'll Miss

This is the harder part to write about.

I'll miss the Myelin team. Building something from near-zero with a small group of people who genuinely care is a rare thing. The late-night debugging sessions, the whiteboard arguments about model architecture, the chai runs when nothing was working.

I'll miss Bangalore food. This is not a small thing. The dosa place near my PG that knows my order. The biryani from Meghana Foods. I've heard Boston has good food, but I have serious doubts about finding a proper filter coffee there.

I'll miss the speed of shipping. In a startup, you can go from idea to deployed model in a week. I know academia moves at a different pace. I'm bracing myself for that.

The Visa Situation

I've started the F-1 visa process and, look, dealing with the US consulate system is its own kind of engineering challenge. Gathering documents, filling forms, preparing for the interview. My mom has a folder that's thicker than my undergrad thesis at this point. Everyone I know who's done this has a horror story. I'm trying not to think about it too much.

Moving Forward

I keep telling myself this is an investment. Two years of depth to build on a decade of breadth. The industry will be there when I come back, and I'll be a fundamentally better engineer for having done this.

But right now, sitting in my Bangalore flat with seven months of packing and goodbyes ahead of me, it just feels like a very big leap. I think that's okay.