mushaf_mughal Profile Banner
Mushaf S. Profile
Mushaf S.

@mushaf_mughal

Followers
29
Following
241
Media
15
Statuses
74

AI/ML Engineer | End Goal is @Nvidia | I talk about everything I’m knowledgeable about | 🦥

Joined June 2021
Don't wanna be here? Send us removal request.
@EXM7777
Machina
2 days
GPT-5.2 is the proof that benchmarks can't be trusted
123
95
2K
@techwith_ram
𝗿𝗮𝗺𝗮𝗸𝗿𝘂𝘀𝗵𝗻𝗮— 𝗲/𝗮𝗰𝗰
3 days
Check out their blogs if you are into AI/ML. 1) Andrej Karpathy Neural networks & LLMs explained from first principles by one of the OGs of modern AI. - https://t.co/LKTDzt0IFA 2) Sebastian Raschka, PhD Deep dives into LLM training and fine-tuning with super clear code
10
176
962
@godofprompt
God of Prompt
5 days
Top engineers at OpenAI, Anthropic, and Google don't prompt like you do. They use 5 techniques that turn mediocre outputs into production-grade results. I spent 3 weeks reverse-engineering their methods. Here's what actually works (steal the prompts + techniques) 👇
41
72
520
@mushaf_mughal
Mushaf S.
8 days
I must say I love Grok for this thing man
@XFreeze
X Freeze
11 days
No other AI answers as boldly as Grok - fearless with the facts, doesn't pull back on uncomfortable truths and designed to be maximally truth-seeking, not censorship
0
0
1
@mushaf_mughal
Mushaf S.
8 days
If China isn’t matching the tech curve, it’s only a matter of time before they leapfrog it. The world should be paying attention. That’s why Taiwan and *TSMC* matter. When one island produces most of the world’s advanced chips, the entire global economy rests on a single point
0
0
1
@Python_Dv
Python Developer
10 days
RAG was supposed to make LLMs smarter. Ground them in facts. Give them memory. But the truth? Most RAG systems today are just fancy search engines—fetching chunks and hoping the model figures it out. That’s not intelligence. The real upgrade is Agentic RAG. Tools like Glean,
38
355
2K
@Hesamation
ℏεsam
9 days
> fine-tune a small LLM > make a reasoning LLM > RL an LLM on a game env > build synthetic data > make a coding agent > build a deep research agent > contribute to an agentic framework these are all hands-on projects that are worth 10 online courses. just code something.
61
198
2K
@googleaidevs
Google AI Developers
11 days
Check out these System Instructions for Gemini 3 Pro that improved performance on various agentic benchmarks by up to ~5%.
95
469
5K
@ChrisLaubAI
Chris Laub
14 days
This Stanford University paper just broke my brain. They just built an AI agent framework that evolves from zero data no human labels, no curated tasks, no demonstrations and it somehow gets better than every existing self-play method. It’s called Agent0: Unleashing
114
401
2K
@rryssf_
Robert Youssef
18 days
This NVIDIA paper just broke my brain. Everyone keeps talking about scaling transformers with bigger clusters and smarter optimizers… meanwhile NVIDIA and Oxford just showed you can train billion-parameter models using evolution strategies a method most people wrote off as
45
331
2K
@mushaf_mughal
Mushaf S.
19 days
I spent almost my entire day testing different SERP tools for an agent I’m building. The whole point of the agent is to pull accurate, fresh info from the internet, so I tried the usual big names. @ExaAILabs says it’s “search built for AI,” @tavilyai promises to connect
1
1
3
@mushaf_mughal
Mushaf S.
19 days
I spent almost my entire day testing different SERP tools for an agent I’m building. The whole point of the agent is to pull accurate, fresh info from the internet, so I tried the usual big names. @ExaAILabs says it’s “search built for AI,” @tavilyai promises to connect
1
1
3
@mushaf_mughal
Mushaf S.
19 days
I spent almost my entire day testing different SERP tools for an agent I’m building. The whole point of the agent is to pull accurate, fresh info from the internet, so I tried the usual big names. @ExaAILabs says it’s “search built for AI,” @tavilyai promises to connect agents to
0
2
3
@mushaf_mughal
Mushaf S.
19 days
I spent almost my entire day testing different SERP tools for an agent I’m building. The whole point of the agent is to pull accurate, fresh info from the internet, so I tried the usual big names. @ExaAILabs says it’s “search built for AI,” @tavilyai promises to connect agents to
0
2
3
@mushaf_mughal
Mushaf S.
20 days
So @Google Gemini 3 toppled all LLMs on the leaderboard. It is now a master in programming, image generation, image editing, and video creation. It took them almost three years to outrank @OpenAI GPT. One thing I liked the most is there are almost no spelling errors in Nano
0
0
1
@akshay_pachaar
Akshay 🚀
22 days
Google just dropped "Attention is all you need (V2)" This paper could solve AI's biggest problem: Catastrophic forgetting. When AI models learn something new, they tend to forget what they previously learned. Humans don't work this way, and now Google Research has a solution.
260
1K
6K
@Shruti_0810
Shruti Codes
23 days
Dear Future AI Engineer, If you want to break into AI in 2025, stop chasing trends. Start mastering the fundamentals. I’m giving away 2 must-read O’Reilly books 📚 that every AI Engineer swears by — from Deep Learning to NLP with Transformers. These books will change your
419
403
2K
@rpnickson
Roberto Nickson
25 days
Meta cooked! 🔥 Their latest Segment Anything models have arrived, SAM 3 and SAM 3D. In this video, I walk you through SAM 3, how it works, why it’s important, and the opportunity it presents for developers, creators and entrepreneurs. The model enables detecting, isolating,
13
45
358