last online: probably debugging something that worked fine yesterday
current status: abusing claude code at 2am and calling it "productivity"
Final year CS student. India. Government college, no fancy placements. Contributing to aden-hive (YC-backed) as an open source contributor. Previously Power BI dev intern at Softude — built dashboards nobody asked for but everyone needed.
Started writing brute force solutions in 2022. Still sometimes write brute force solutions in 2026. Some things don't change.
I don't have a mentor. I have Stack Overflow, a lot of failed submissions, and a very patient terminal.
current_state = {
"learning": ["PySpark", "Airflow DAGs that don't randomly die", "system design"],
"grinding": "CF + LC daily — Sahi01 on Codeforces",
"building": "distributed workflow engine (mini Airflow clone, Node.js + MongoDB)",
"applying": ["MLH Fellowship Summer 2026", "LFX Mentorship Term 2", "Outreachy"],
"OS": "Fedora Linux — yes it matters, no I won't explain",
"tool_abuse": "Claude Code for everything. no regrets.",
"sleep": None # pending
}tried a PR today. it's 11pm. probably not getting merged this month either.
| frequency | reality |
|---|---|
| try daily | ✅ open issues, read code, cry a little |
| closed monthly | if I'm lucky and the maintainer is awake |
| merged yearly | it happened. I have proof. |
Currently contributing to aden-hive and picking up issues at TheAlgorithms, HuggingFace Transformers, and whatever YC repo has a good first issue open at midnight.
The strategy: find small repos before they get famous. Merge fast. Repeat.
Languages I actually use
Data stuff I'm actually learning (not just listing)
ML / AI (when the math makes sense)
Tools I use daily without thinking
Codeforces handle: Sahi01
grinding since 2022. started with brute force. still sometimes submit brute force at contest end when the clock is at 0:02. no shame.
topics I can actually solve vs topics I tell people I know:
can solve → greedy, basic DP, graphs (BFS/DFS), binary search, prefix sums
sometimes → segment trees, bitmask DP, flows (if I've seen it before)
actively → number theory, advanced DP, tree DP (getting there)
lying about → nothing, I stopped doing that in 2023
I actually like math. Not "puts it on resume" like. Genuinely spent a weekend on Bayesian inference and didn't hate it. Read about eigenvalues because PCA stopped making sense. That kind of like.
- probability & stats — for ML that isn't just copying a kaggle notebook
- combinatorics — for CP problems that look simple until they aren't
- linear algebra — SVD and PCA finally clicked after the third try
- calculus — gradient descent from scratch once, never again, but I get it now
- Distributed Workflow Orchestration Engine — mini Airflow in Node.js. Kahn's algo for DAG scheduling, isolated-vm for sandboxing, SSE for real-time logs. It's cursed and I love it.
- LFX Mentorship Term 2 — pre-application contributions in progress
- MLH Fellowship Summer 2026 — essays done, fingers crossed
- recovering my X account (@TheBlackHerald7) — suspended, appealing, will update
wake up → check CF → solve something → commit something → break something → fix it at 2am with Claude Code → sleep → repeat
some days the PR gets merged. most days it doesn't. that's fine. the repo is still there tomorrow.




