Tushar Vaghela

Director of Engineering @proxwrks 🚀 | Reader 📚 | Learner 🧗🏻‍♂️ | Documenter 📒 | | | |

[Tushar Vaghela @tv1981]

So on Everest Day, celebrate the climbers, but also the surveyors, Sherpas, porters, families, and silent believers behind every “solo” win. ❤️🏔️

#EverestDay

[Tushar Vaghela @tv1981]

Mount Everest is the highest point on Earth. But its greatest lesson may be deeply human: the journey is what stretches our limits, softens our ego, and shows us who we become on the way up. 🏔️✨

#InternationalEverestDay #MountEverest

[Tushar Vaghela @tv1981]

1921 explored the northern approaches, 1950-51 made the workable south-side path clear, and on 29 May 1953 Tenzing Norgay and Edmund Hillary made the first confirmed summit. 🙌

[Sameer S @sameergshah]

There’s a very small window right now where tiny teams can build things that look unfair.

We’re in that window.

Hiring AI Engineers - Mumbai
•⁠ ⁠freshers
•⁠ ⁠interns

If you care about product, speed, and building AI-native experiences instead of AI gimmicks — DM me

[Sunny R Gupta 🐰 @sunnykgupta]

I knew someone who always had the next big thing.

Every time we met:
- New idea.
- New plan.
- New life blueprint.

And honestly?
They sounded convincing every single time.

[svs 🇮🇳 @_svs_]

come by if you can! i'll be showing off my brand new personal agent - karinda - and talking about why elixir is the right choice for making agents.

https://t.co/qq5E1JFPPO

[Sunny R Gupta 🐰 @sunnykgupta]

A lot of very good engineers secretly feel they’re already behind on AI coding tools.

Not because they haven’t tried Claude Code, Cursor, or Codex…
but because everywhere they look, there’s some mysterious combo of rules, hooks, skills and “best practices” they’re supposed to know.

So everyone does the same thing in isolation:
- Rewriting AGENTS.md, CLAUDE.md, .cursor/rules in each repo
- Learning the same painful lessons on security, git and testing
- And quietly wondering if they’re “using these tools wrong”

I wanted to fix a small but real part of that.

Over the last days I pulled together a set of opinionated, battle-tested rule files for:
- Claude Code → CLAUDE.md
- Codex → AGENTS.md
- Cursor → .cursor/rules/*.mdc

They focus on the stuff that actually changes how your agents behave:
🔒 Clear security boundaries (what the agent must NEVER do)
🧠 Code style that encodes real architectural decisions, not generic lint rules
🧪 Testing philosophy (when not to add tests, and why)
🔀 Git & workflow rules that stop agents from doing surprising repo-wide edits

No threads. No PDFs.
No “comment Interested to get the link”.

The GitHub repo link is in the replies so you can just open → copy → adapt → ship.

[Sunny R Gupta 🐰 @sunnykgupta]

1/ Most people “prepare for interviews”.

If you’re aiming for top spots like @Rippling , @databricks , @stripe , @nvidia , @Razorpay , @JioHotstar etc, you may need to rebuild yourself for that bar.

Here’s the PLAYBOOK for 0–5 YOE folks in India.
Save this. 🔖

[Tushar Vaghela @tv1981]

@sunnykgupta Hello 🙏🏻

[Andrej Karpathy @karpathy]

Software horror: litellm PyPI supply chain attack.

Simple `pip install litellm` was enough to exfiltrate SSH keys, AWS/GCP/Azure creds, Kubernetes configs, git credentials, env vars (all your API keys), shell history, crypto wallets, SSL private keys, CI/CD secrets, database passwords.

LiteLLM itself has 97 million downloads per month which is already terrible, but much worse, the contagion spreads to any project that depends on litellm. For example, if you did `pip install dspy` (which depended on litellm>=1.64.0), you'd also be pwnd. Same for any other large project that depended on litellm.

Afaict the poisoned version was up for only less than ~1 hour. The attack had a bug which led to its discovery - Callum McMahon was using an MCP plugin inside Cursor that pulled in litellm as a transitive dependency. When litellm 1.82.8 installed, their machine ran out of RAM and crashed. So if the attacker didn't vibe code this attack it could have been undetected for many days or weeks.

Supply chain attacks like this are basically the scariest thing imaginable in modern software. Every time you install any depedency you could be pulling in a poisoned package anywhere deep inside its entire depedency tree. This is especially risky with large projects that might have lots and lots of dependencies. The credentials that do get stolen in each attack can then be used to take over more accounts and compromise more packages.

Classical software engineering would have you believe that dependencies are good (we're building pyramids from bricks), but imo this has to be re-evaluated, and it's why I've been so growingly averse to them, preferring to use LLMs to "yoink" functionality when it's simple enough and possible.

[Sunny R Gupta 🐰 @sunnykgupta]

🧵 If you’re in your early 20s, hungry, and actively looking for hustle culture (with real upside, not fake “startup vibes”), here’s a 2026 list of companies I’d aim for from India.

Remember, this isn’t a “safe” or “complete” list ;)

🔖 Bookmark for later

[Hyperterse @hyperterse]

Give your agents structured access to your databases in minutes, not months, and make them work for you.

Here’s a quick demo of what Hyperterse can do for you ⬇️

[Saket Tawde @SaketCodes]

Want to use your coding skills to make a real difference?

We're a nonprofit hiring fullstack devs — the mission matters more than the perks.

No flashy benefits package — just honest pay and genuinely fascinating problems to solve.

https://t.co/CHQ3YaXT2Q