Ayush

Crafting objective proof from vast and ambiguous datasets. Validating a chosen path to enable confident, decisive action.

#100DaysOfML

[Ayush @TensorThrottleX]

Embarking on a 100-Day Challenge. Starting from 30th December.
This time, it's fundamentally different. We @TensorThrottleX, @BinaryBlaze16, @CodeAyushD committing to unflinching transparency, no curated highlights, no polished outcomes. Only the raw work itself.

[DIVYANSH @DsThakurRawat]

Hello
@X
@nikitabier, Yesterday, one of my friend's Twitter accounts @rrrautela

got suspended without any reason. He only posts about his DSA journey and job hunting. He doesn’t do any shitposting, yet his account still got suspended.

[Ayush @TensorThrottleX]

@HarshalsinghCN Working on something kinda similar rn, gonna check it out, lowkey need something like this myself😅

[Ayush @TensorThrottleX]

@Pseudo_Sid26 Man that's really good, excited for what's to come..........

[Siddharth @Pseudo_Sid26]

https://t.co/uByNB90dLR

[Ayush @TensorThrottleX]

@dhruvtwt_ Damn bro......so fucking cool.

[Ayush @TensorThrottleX]

@dhruvtwt_ @nvidia Absolute🤞

[Dhruv @dhruvtwt_]

Why is no one talking about this?

@nvidia is offering around 80 AI models via hosted APIs absolutely for free.

You get access to MiniMax M2.7, GLM 5.1, Kimi 2.5, DeepSeek 3.2, GPT-OSS-120B, Sarvam-M etc.

This plugs straight into OpenClaude, OpenCode, Zed IDE, Hermes agent and even with Cursor IDE.

Setup:
– Grab API key: https://t.co/Wfdclm0hY2
– base_url = "https://t.co/VOGC10LmGP"
– api_key = "$NVIDIA_API_KEY"
– select model (e.g. minimaxai/minimax-m2.7)

If you’re building or experimenting, this is basically free inference.

Lock in and start building today anon.

Thank me later.

[Ayush @TensorThrottleX]

@LowKeyDevs I had this idea earlier(like thoughts, manifold, projects, yaps )and even built it to a certain level. I also thought of pairing it with Notion for a more dynamic and effective output, but in the end,
over-optimization killed it.

[Ayush @TensorThrottleX]

@Vixhal Farming

[Ayush @TensorThrottleX]

@Vixhal Clarity, but my thoughts tend to spiral into overthinking loops.

[Ayush @TensorThrottleX]

@Sudipto737 That's a great initiative, keep going. And ur post isn't popping up on my feed.

[Ayush @TensorThrottleX]

@lochan_twt Yeah bro, you’re the “Spidey head” for that AI-ML sheet.

[Ayush @TensorThrottleX]

@TheByteDax Go for it brother 🤞

[Ayush @TensorThrottleX]

Been a little off the grid for the past few days due to some health stuff and a few things on the side.
Will be back in the flow, soon.
till then,
peace out🤞

[aditya @adxtyahq]

claude opus 4.7 just dropped
here’s how to cut your token usage by ~60%

[Claude @claudeai]

Introducing Claude Opus 4.7, our most capable Opus model yet.

It handles long-running tasks with more rigor, follows instructions more precisely, and verifies its own outputs before reporting back.

You can hand off your hardest work with less supervision.

[Ayush @TensorThrottleX]

@idkwhyvi62159 Was the interview purely for the ai role? Or something else?

[Devaansh Bhandari @ThisIsBhandari]

If you're a tech student grinding every single day, listen to this.

My 4 years of college felt like pure torture.

I had 8 backlogs, no campus placements, no DSA.
That’s when everyone wrote me off.

Profs called me useless.
Peers treated me like a failure.

I slept with anxiety.
Woke up with doubt.

It felt like a massive gamble —
either I build something big
or end up with nothing.

But I kept going anyway.

Late night coding sessions.
Missed parties.
Showing up even when I didn’t want to.

Somehow… it paid off.

Today I’m a founding engineer at a YC startup,
onboarding clients worth millions in MRR.

If you’re in the messy middle right now,
don’t stop.

One opportunity can change everything.❤️

[Ayush @TensorThrottleX]

@like_backspace Enjoy the moment bro🤞

[Ayush @TensorThrottleX]

Day 100 :

[Ayush @TensorThrottleX]

Day 317 : DataScience Journey
understanding bagging and RF, the next is boosting & the shift in thinking here is really interesting. While bagging works by training multiple trees independently and avg. their results, boosting does the opposite: it builds models sequentially,