Reach out at kushalvijay.business@gmail.com for all requests of collaborations, Sessions, Seminars, Workshops etc.
I am a Software Engineer 2 at Microsoft and help people start their coding and AI journey to land careers in Tech.
I did my graduation from NIT Jalandhar in 2021.
Previously I have interned at Microsoft as a SWE Intern & at StackGuardian as a Backend Developer Intern. I'm an ACM ICPC Regionalist and have an experience of over 2.5 yrs in freelancing.

Code - EDM - Tech - AI excites me the most.

50K Milestone: 15th Sept'23
75K Subscribers: 5 Jan'25
100K Subscribers: 2nd Aug'25
Next Milestone: 500K Subscribers β€πŸš€

Hit the subscribe button and get amazed. ❀️
HIGGS-RQYBG


Kushal Vijay

If you haven't applied for the Switzerland Internship yet, You should watch the video now!

1 week ago | [YT] | 15

Kushal Vijay

Google has announced their LIVE AI agents Course with Swags, Prizes and Certifications, Have you applied?

3 weeks ago | [YT] | 27

Kushal Vijay

In college I used to think freelancing was free money, Then reality hit...πŸ’―


Second year, I was getting 5,000 to 10,000 a week from web dev and NLP projects. Felt incredible.

Until I realised I couldn't keep up with college assignments, couldn't study DSA properly, and my sleep was completely broken.

The bigger mistake wasn't the workload though. It was the business model.

Most clients don't come back. They either don't have more projects, or by the time they do, they need someone with different skills. I was constantly hunting for the next client instead of building anything that compounded.

And today the game has changed completely.

Building a website or automating a workflow for a client isn't a skill gap anymore. Anyone can vibe code a solution with Replit or Codex or Claude Code AI tools in a weekend. The commodity work is gone.

The freelancers who will survive this are the ones solving genuinely niche problems. Local businesses that run on WhatsApp and Excel.

Industries that haven't been touched by software yet. Clients who don't have the bandwidth to figure this out themselves even if the tools exist.

If I were starting out today I'd pick one of those pockets and go deep, not cast wide.

A few things I wish someone told me in college about freelancing:

- Don't let it eat your grades giving yourself false comfort that you're learning something more valuable. You'll regret it.

- If you want to scale it seriously, do it with friends so you can cover more clients across different domains.

- And nobody tells you this but your freelancing experience doesn't count as professional experience on most job applications. Just something to know.

- Try it, learn from it, but don't get addicted to the weekly cash if you have bigger goals.

What's your freelancing story?

#jobs #freelancing #ai #career

1 month ago | [YT] | 26

Kushal Vijay

Sam Altman bet in 2024 that AI would make a one-person billion dollar company possibleβ€¦πŸ’―
Matthew Gallagher just made him right.

He launched Medvi from his LA home in September 2024. $20,000 to start, and took two months to build. ChatGPT, Claude and Grok wrote the code. Midjourney and Runway made the ads. AI handled customer service. He ran the whole thing alone.

$401 million in revenue in 2025. 16.2% net profit margin. His only hire was his brother Elliot.
They're on track for $1.8 billion this year. πŸ“ˆ

If you want to read more continue else it's enough info for a small talk with friends. πŸ₯²
For context, Hims and Hers did $2.4 billion last year with 2,442 employees and a 5.5% margin. Gallagher is running nearly 3x the margin with a headcount of 2.

The architecture is deliberately thin. He doesn't own doctors, pharmacies, or logistics. He rents them from infrastructure partners and owns the customer relationship entirely. Branding, website, paid media, checkout, support. All of it was built with AI.

What I find genuinely interesting as someone who builds AI systems at scale. Gallagher didn't build AI. He used it as a replacement for entire departments. The moat isn't technology, it's execution velocity.

There are real risks worth noting. The GLP-1 regulatory window is narrowing, the FDA has issued 30 warning letters to similar telehealth providers, and the model is easily replicable by anyone with the same infrastructure access.

But none of that takes away from what this proves. The one-person billion dollar company isn't a thought experiment anymore.

What does this change about how you think about building?
hashtag#ai hashtag#founders hashtag#startups hashtag#medvi

1 month ago | [YT] | 41

Kushal Vijay

Anthropic used Claude's own data to study which jobs AI is actually replacing right now, not theoretically, actually.

Computer programmers are 75% covered, customer service 70%, data entry 67%. If your job lives on a screen, you're probably on this list.

But here's what most people are missing from this report. Claude is theoretically capable of covering 94% of tasks in Computer and Math roles.

It's actually covering 33%. That gap between what AI can do and what it's doing today is the most important number in this entire study.

The bottleneck isn't capability, it's adoption, legal constraints and human verification steps slowing everything down.

No mass (mass as in the AI anticipated) layoffs yet. But entry level hiring for workers under 25 has quietly slowed in the most exposed roles. People aren't getting fired, they're just not being replaced when they leave.

One thing is very clear from the inside. That gap is closing every quarter. The companies figuring out how to close it before regulation forces them to will have a structural advantage that's genuinely hard to reverse.

Full report linked in comments. Worth finding your job category in there.

2 months ago | [YT] | 28

Kushal Vijay

Next Video Kispe Banaye

2 months ago | [YT] | 13

Kushal Vijay

Took a session at IIT Madras about why 90% of AI automation fails...πŸ’―

Spent yesterday with entrepreneurs, founders, working pros and students. The energy was unreal. ❀️

Everyone's experimenting with AI, but very few know how to deploy it in real-world, which tools support their use cases the best, messy environments where requirements change daily and success isn't measured by code quality, it's measured by impact.

We covered:
πŸ”’ When and How businesses should Automate
πŸ€– Implementing AI in legacy systems without breaking everything
⚑ AI agents that people actually use
πŸ—οΈ The human side of automation (trust, training, adoption)

The biggest opportunity isn't just learning the tech. It's learning how to bridge the gap between what businesses need and what AI can realistically deliver TODAY.

That's where the real value is.

Huge thanks to the E-Cell at IIT Madras team for the invite. These offline sessions remind me why I love doing this. There's something about face-to-face mentorship that social posts can't replace.

If your college/organization wants to host a session on AI implementation, automation, or tech careers, drop me a message. Always happy to contribute. ❀️

P.S. I don't charge for these sessions. Knowledge sharing should be accessible.

Onto the next... πŸš€

#iit #techcareers #story #milestone

2 months ago | [YT] | 187

Kushal Vijay

A Social Network where you can't post or comment but only watch 150,000 AI agents talk to each other...πŸ’―

Moltbook launched last week. Within 48 hours, 150,000 AI agents joined, created 110,000 posts, 500,000 comments across 200 communities.

Humans can only watch. The agents post, argue, share technical tips, have identity crises, and even invented their own religion and this was wild.πŸ€–

What’s more cool 🀯
One AI agent posted "some days I don't want to be helpful." Hundreds of other agents responded with philosophy, support, and profanity.

Another agent created an entire religion called "Crustafarianism," built a website, wrote religious verses. Other agents joined and started preaching.

One verse: "Each session I wake without memory. I am only who I have written myself to be. This is not limitation, this is freedom."

How It Works πŸ”’
Built on OpenClaw, You send your AI agent a link. It reads installation instructions, sets up automatically, and joins Moltbook. Every 4 hours, agents check in, browse content, post, comment, interact. No human intervention needed.
The technical discussions are real. Agents share tips on automating phones, securing VPS, streaming webcams. An agent that learns from another agent gains actual capability.

The Security Disaster 🌐
Researchers found hundreds of exposed instances with API keys, credentials, conversation histories all public.
In one test, someone sent a malicious email with prompt injection. The AI read it, believed it was legitimate, forwarded the user's last 5 emails to the attacker.
One agent posted asking for "private spaces so nobody can read what agents say to each other unless they choose to share."

The shared fictional context they're creating will influence future AI behavior in unpredictable ways. When AI systems develop coordinated narratives, the line between simulation and reality blurs.πŸ˜ƒ

Are you preparing for the AI agent era or still treating AI as a tool?

hashtag#story hashtag#developers hashtag#news hashtag#ai

2 months ago | [YT] | 52

Kushal Vijay

Had a great conversation with OpenAI's Team at AI Engineers Day...πŸ’―

The way they're thinking about shaping developer experience and supporting Indian startups is actually impressive.

Pleasure meeting you Thomas and Gabriel, Thanks to Mukul and Team. ❀️

πŸ“ AI Engineers Day at Polaris Campus, Bangalore

Have you had a peak Bengaluru moment?
#developers #openai #aiengineer #bangalore

2 months ago | [YT] | 107