By Vivek Kumar Verma
I keep meaning to write one of these "what I'm up to" posts, and I keep putting it off because there's always something half-finished that feels more urgent. So this is me stopping for a minute to take stock. Some of this is going well, some of it is messy, and I'd rather show both.
I'm in my final year of B.Tech in Computer Science at GGSIPU, graduating in 2027, living in New Delhi. I work mostly with the MERN stack, and I've been freelancing since 2024. Last year I did a full-stack internship at MAHIRAG, and this year I spent March to September on a remote internship with Samikaran India. That one just wrapped up in early September, so a lot of my head space is still there.
Samikaran is an NGO, and the work was real in a way college projects rarely are. I built out their donation flow with Razorpay, used Supabase as the database, and put together a CMS so the team could look at transactions and edit the content on their own site without pinging a developer every time.
Payments are the part that humbled me. When money is involved, "it works on my machine" stops being an acceptable answer. You start thinking about what happens when a payment succeeds but the page never loads, or when someone refreshes at the wrong moment. I'm a more careful engineer than I was in March, and I'd take that over any certificate.
My team, Problem2Product, is working on MEDHAVI for SIH 2026. The problem statement is from the Ministry of Tribal Affairs: an AI-enabled scholarship and fellowship management system for Scheduled Tribes students. If you've ever seen how scattered and paper-heavy scholarship processes can be, you know why this one is worth attempting.
The plan is Next.js with TypeScript and Tailwind on the front, Node.js and Supabase behind it, OCR for reading documents, and DigiLocker integration so students aren't uploading the same papers again and again. Honestly, the hard part isn't the stack. It's designing something that works for people with patchy internet and little patience for confusing forms. I keep reminding myself that the user here is not another developer.
PackLM is an open-source, token-efficient structured data format for LLMs. It's on PyPI and gets somewhere around 100 to 200 monthly users, which is small by internet standards but wild to me, because those are strangers choosing to install something I wrote. I've shipped v1 and v2 and I'm still thinking about where it goes next.
Slop Scan is the one I enjoy most. It's a React tool that looks at a repo or a live site and scores how much of it looks AI-generated or vibe-coded, using 18 heuristic rules, and then roasts you about it in plain language. I built it partly as a joke and partly because I genuinely care about this. I'd like the things I make to feel like a person made them.
A Bitcask-style key-value engine is my "go deeper" project. I wanted to understand storage by building it instead of just reading about it, so I've been working through it in phases. It's slow, it's unglamorous, and I'm learning more from it than from most tutorials.
Gridlock City started as a simple Three.js car game and kept growing into a free-roam city with traffic, collisions and a minimap. No career purpose at all. It's just fun, and I think every developer needs one of those.
I'm running a 100 Days of DSA challenge on YouTube and a series called Fifty Shades of JavaScript on Instagram and YouTube. Teaching something the day after I learn it is the fastest way I know to find out what I only half understand. Posting daily is also tiring, and some days I'm very aware of how few people are watching. I keep going anyway.
I'm looking for frontend and full-stack roles, in India or remote internationally. It's the least fun part of this list. Tailoring applications, rewriting my resume, sending LinkedIn messages that I hope don't sound like every other message in someone's inbox. One rule I've held onto: I'm not going to claim things I haven't done. A resume that's honest and a bit shorter beats one that falls apart in the first interview question.
Ship MEDHAVI's submission, keep the DSA series going, land the right role, and finish the things in my graveyard of half-built projects, or at least admit which ones I'm actually abandoning.
If you're building something similar, or you're hiring, or you just want to argue about whether a project needs a database at all, I'd like to hear from you.
