SHIVAM KUMAR

AI/ML • DevOps • Inference Engineering • BITS Pilani

[SHIVAM KUMAR @shivam2003_dev]

Cloud CPU & Memory Benchmarks, 2026 — who really wins per vCPU, per GB, per dollar. ?
read : https://t.co/wsvNVpPVOp

[SHIVAM KUMAR @shivam2003_dev]

Claude usage limits are frustrating.
I was at ~60% usage, ran /compact, and my quota jumped to 100% almost instantly. After the reset, even loading previous context consumed a huge portion of the available budget before doing any real work
@AnthropicAI @ClaudeDevs @claudeai

[Matt @matt_dz]

@0xAX Andrej Karpathy's videos, https://t.co/72J0Bg4rwD
Sebastian Raschka's, https://t.co/rBUMiY3IEt
Stanford:
CME295: LLMs, https://t.co/PRYD0Hz72X
CS25: Transformers United, https://t.co/yvHbM6cMXh
CS336: Language Modeling from Scratch, https://t.co/NoeyQlbBri.

[Dan Kornas @DanKornas]

🔗 GitHub: https://t.co/gmMc7QE3Fs
---
✉️ If you’re into AI, ML, agents, and building real systems, join my newsletter (it’s free): https://t.co/zJ9uwd6qSd

[SHIVAM KUMAR @shivam2003_dev]

Learning streak in full swing! https://t.co/EO9zMxatnf #tryhackme via @tryhackme

[SHIVAM KUMAR @shivam2003_dev]

I just completed Zero Logon room on TryHackMe! Learn about and exploit the ZeroLogon vulnerability that allows an attacker to go from Zero to Domain Admin without any valid credentials. https://t.co/FMJIGIbN9Y #tryhackme via @tryhackme

[Ada Fang @AdaFang_]

AI Scientists are starting to actually do science. Not just answer questions. Not just run workflows.

Introducing AutoScientists: a decentralized team of AI agents that can generate hypotheses, design experiments, write code, test ideas, analyze failures, and revise strategy as evidence accumulates.

Because real research is not a to do list of tasks.

It is a living search process. Leads emerge, failures matter, teams form around what works, and priorities shift when evidence changes. Much like how a lab of scientists would work on cutting edge research together.

Across GPT training optimization, biomedical ML, and protein fitness prediction, this decentralized structure consistently does better research.

Learn more 👇

@GaoShanghua @marinkazitnik @KempnerInst @HarvardDBMI @Harvard

[SHIVAM KUMAR @shivam2003_dev]

I just completed AI Models & Data room on TryHackMe! Explore how data is fundamental to AI security, and the models which power it. https://t.co/oFa0geA9QL #tryhackme via @tryhackme

[SHIVAM KUMAR @shivam2003_dev]

I just completed AI/ML Security Threats room on TryHackMe! Learn AI basics, key terms, and how it's used by both attackers and defenders. https://t.co/G2gXG8tSDH #tryhackme via @tryhackme

[SHIVAM KUMAR @shivam2003_dev]

I just completed AI Threat Modelling room on TryHackMe! Assess and mitigate enterprise AI/ML risks via systematic, defender-focused auditing. https://t.co/cD5w8rejhI #tryhackme via @tryhackme

[SHIVAM KUMAR @shivam2003_dev]

Learning streak in full swing! https://t.co/hukLDGJWWw #tryhackme via @tryhackme

[SHIVAM KUMAR @shivam2003_dev]

"AI systems are no longer in the pilot phase. Language models handle customer support queues. Fraud detection engines make real-time authorization decisions affecting millions of transactions daily. Recommendation systems" https://t.co/xBPiXPeWMv

[SHIVAM KUMAR @shivam2003_dev]

https://t.co/4HGb5ai0uR #tryhackme via @tryhackme

[SHIVAM KUMAR @shivam2003_dev]

I just completed Supply Chain Attack Vectors room on TryHackMe! Learn how trusted ML components can be turned into attack vectors. https://t.co/Qw7wNHYOHr #tryhackme via @tryhackme

[SHIVAM KUMAR @shivam2003_dev]

I just completed Understanding AI Supply Chains room on TryHackMe! Explore how AI's dependency on external models, datasets, and packages creates attack surfaces. https://t.co/mo9qzjzU8a #tryhackme via @tryhackme

[SHIVAM KUMAR @shivam2003_dev]

Calculators would “destroy math skills”
Search engines would “destroy memory”
IDE autocomplete would “destroy programming”
Compilers would “destroy understanding”
High-level languages would “destroy real engineering”
is it ?

[SHIVAM KUMAR @shivam2003_dev]

"If you're me (a DevOps + ML apprentice): Build the skills that orchestrate AI rather than compete with it. Learn the infrastructure layer. Learn how models are deployed, evaluated, monitored, and integrated with the rest" https://t.co/1latVUi0Xe

[SHIVAM KUMAR @shivam2003_dev]

Will AI Destroy Jobs or Create Prosperity? — What the Math Actually Says. ?
https://t.co/1latVUi0Xe

[SHIVAM KUMAR @shivam2003_dev]

"want your AI to count flamingos in a photo. The usual approach? Collect 50,000 labeled images of flamingos. Annotate each one by hand. Train a model for three days. Pray it generalizes. Then repeat the entire process whe" https://t.co/3NeffUhF4r

[SHIVAM KUMAR @shivam2003_dev]

"Researchers at Google Research, Google DeepMind, and NYU asked: what if we could compress those vectors to 2–4 bits per number without losing the math that matters? Their answer is TurboQuant, and the core trick is almos" https://t.co/sYY0uts73k