
Founder, Futurense | Co-founder, 1% Club @FDE_Academy - World’s First Structured Prog
Joined September 2018
- Tweets775
- Following255
- Followers1.6K
- Likes453
Pinned Tweet
I had a talk with @BeerBicepsGuy about:
- work study balance
- Will AI get banned?
- The future of the tech game in india
- India’s GDP
- How companies are going to use AI
And SO MUCH MORE.
Check it out: youtube.com/watch?v=tZ6uG_SY…
The internet cancelled OpenAI’s new model, but India’s back offices desperately needed it.
Reddit users shrugged at GPT-6.1 Terra. Some wanted something stronger than Claude Opus 5.5. OpenAI’s own Ted Sanders told Hacker News, “For subscribers, it’s not a bombshell release.”
The independent numbers show the trade-off. Artificial Analysis scores Opus 5.5 at 58 and Terra at 48.
It puts Opus at $5.98 a task and Terra at $0.40.
Greg Brockman says Terra uses a computer faster than he can. OpenAI’s own example is reading invoices and typing them into accounting software.
And India runs on that kind of work. A back office that clears invoices all day will pick the lower bill.
The model people call average may reach us before the one everyone is waiting for.
78
For an Indian engineer in 2026, moving abroad has started to become a risky bet than building a career in India.
For 30 years it was the opposite. Get the degree, the visa, and get on the flight was the right plan, because the biggest opportunities were never in India.
Now, just to be clear: Indians left because ambition goes wherever it can get the most out of itself.
But look at what happened to the "better way out":
- US student visas issued to Indians in June and July fell 69% in a year, from 41,336 to 12,776.
- H-1B registrations dropped 38.5% in the same cycle, from 343,981 to 211,600. Indians are about 71% of everyone approved on that visa, so this was mostly our story.
- The big IT firms felt it the most. Infosys went from 8,886 registrations to 759 and TCS went from 5,955 to 284.
The home option looks very different from ten years ago.
India’s global capability centres now employ 2.36 million people, more than TCS, Infosys, HCLTech, Wipro and Tech Mahindra combined.
And the pay has followed too.
At these centres a GenAI engineer starts at about Rs 13 lakh, an AI architect with 15 years of experience earns up to Rs 1.08 crore, and a principal data scientist up to Rs 93 lakh.
A career that used to need a visa and a flight can now be built in Bengaluru, Hyderabad or Pune.
The old plan rested on one belief: the work that mattered was somewhere else. The visa route that carried a generation of IT careers is shrinking, and 2.36 million GCC jobs now sit at home.
For an engineer starting out today, I would build the first few years here. Going abroad stays a good option for later, with experience that travels.
If you are still planning the flight, what do you need abroad that Bengaluru, Hyderabad or Pune cannot give you?
121
Indeed says FDE job postings in India grew nearly 19x in a year.
Glassdoor shows many professionals still asking what the role is.
The headline reads as a hiring boom. I keep coming back to the definition gap.
Here is what that gap shows:
Postings began accelerating in March 2026. After the June peak, they ran nearly 16x last year’s level in July and over 6x in August.
Glassdoor threads ask how the role differs from software engineering, which skills to prepare, and where it leads. The postings show why those are hard to answer.
→ Palantir describes its forward deployed software engineer as one customer with many capabilities, and its software engineer as one capability for many customers. Glassdoor’s listings use the title across enterprise AI, data engineering, infrastructure, and healthcare.
→ Glassdoor’s listings run from Python or Java and data engineering to generative AI, agents, cloud, and APIs. Two Palantir engineers add technical decomposition and user empathy.
→ Even OpenAI set up its forward-deployed team in 2025, planning about 50 engineers, per the Financial Times. In May 2026, it agreed to acquire Tomoro, a firm with about 150 forward deployed engineers and deployment specialists.
Indeed calls its FDE sample too small for a reliable median salary.
If an FDE is one customer with many capabilities, I’d expect training to begin with one customer’s problem. That makes this look like a training question first.
Which part of this role is least clear today: the work, the skills, or the career path?
7
242
HCLTech now has 230,339 employees.
It added 6,450 in a single quarter, the most in 16 quarters.
Infosys plans to hire 20,000 freshers this year. Wipro: about 7,500.
Three giants, three different hiring plans.
Here’s what the numbers say:
HCLTech and Wipro employ within 6,000 people of each other. In the last six months, HCLTech onboarded 7,192 freshers and Wipro about 1,000.
Their stated plans point to a different second halfy. If both reach their yearly numbers, Wipro would need to onboard roughly 6,500 more by March, and HCLTech roughly 3,000 to 4,000.
→ Base: Infosys’s plan is about 6% of its 3,32,000+ people. HCLTech’s is 4% to 5% of its workforce, and Wipro’s about 3% of 236,328. Relative to size, Wipro plans a lower intake than the other two.
→ Calendar. HCLTech has used about two-thirds of its year and Infosys more than half. Wipro has used about 13% and says the rest leans toward Q4.
→ Business backdrop: HCLTech reported $3.26 billion in new deals and raised the floor of its FY27 growth guidance from 1% to 2.5%. Wipro’s large deal bookings fell 58.6% sequentially, and headcount fell by 6,716.
Infosys cut its FY27 growth guidance to 1% to 2% and kept its fresher plan.
Each number is a forecast. Infosys held its plan while cutting its outlook, and HCLTech went from 1,056 freshers in one quarter to 6,136 in the next.
HCLTech’s net addition already absorbs exits, with voluntary attrition at 12.8%. Completed onboarding should not be read as a count of current vacancies.
Two of the three have more than half of their intake behind them. Wipro’s December and March quarters will show whether its calendar catches up.
When you plan your own fresher intake, what do you anchor it to: a fixed number, the order book, or the bench?
1
175
Accenture has trained 550,000 employees in generative AI.
Now it funds only $29 a month per employee for Claude.
Here’s what that gap tells you:
Its AI and data team has grown from 40,000 to nearly 110,000 people since 2023. Per 404 Media, some staff had been spending $500 to $1,000 a month on Claude.
→ Since October 1, usage above $29 is charged to a specific project or team through a Work Breakdown Structure (WBS) code.
→ One employee, building an AI accelerator for several Accenture clients, said their usage ran $500 to $1,000 a month.
→ Claude Enterprise stays available based on role and business need. One employee said their access was revoked.
Look at where the decision moved. The cost of heavy AI use now sits on a project’s books, so the person running the project decides which prompts are worth paying for.
I keep coming back to the gap between the two numbers. 550,000 people learned how to use AI. The new rule asks each of them to show what that use returns. A $1,000 bill is cheap if it replaces a week of an engineer’s work, and expensive if it replaces nothing.
For India’s IT services firms, this lands harder. They sell skilled people’s time. If AI cost sits on every project, the engineers worth staffing are the ones who can show what their AI use returns.
📍 $29 is the amount Accenture funds centrally. Use above it can continue if a project pays, and the reporting does not say how many employees are affected.
The old employee came with a salary and a laptop.
The AI-era one comes with a usage bill that someone has to own.
If a project had to pay for your AI usage next month, what would you point to as the return?
Raghav Gupta retweeted
This week felt like the beginning of the reversal of International investor sentiment on India :
> Blackstone COO was here and announced their $100B investment target in India
> Trump waived off the russian oil sanctions
> Murmurs of the US trade deal finally getting signed
> Fairfax bought 60% of IDBI for $5.6B
> the AI trade is showing some signs of cracking with coreweave debt spiking , Anthropic ARR slightly off
> India looking good as an AI hedge , well diversified large economy with booming consumption and manufacturing
People short India should ask themselves if we are growing at 8% with oil at $100 and two global wars, what will India do when the macro turns.
LFG 🚀
The IMF says 40% of the world’s jobs could change or disappear.
Infosys’s chairman calls AI job-loss fears “hogwash.”
Zoho cancelled appraisals this week and is working to avoid mass layoffs.
Here’s what he sees that Infosys’s chairman doesn’t:
Nilekani is looking at demand. TCS’s annualised AI revenue reached $3.1 billion, so the work is coming. He has a point.
Vembu is looking at the bill. The server prices Zoho depends on are up 4x, its headcount hasn’t grown in 18 months, and he’s said the money meant for new hires is now going to AI and data centres.
Many companies now have one budget and two bidders. A person and an AI.
Hiring is already shifting. 64% of new jobs in India’s global capability centres this year require AI, data or automation skills. TCS says about 80% of its future hires will have digital and AI skills.
TCS is also building a team of up to 8,900 forward-deployed engineers, a role OpenAI and Anthropic are hiring for too.
Somewhere right now, an engineer with your degree is learning to deploy AI inside a real company. Their resume will land on the same hiring manager’s desk as yours.
Whoever is right, hiring is already tilting toward people who can make AI work.
Companies will keep choosing between a person and an AI. Make sure they pick you.
111
In 1961, India's top 1% owned about 13% of India's wealth.
Today, they hold around 40% of it, per the World Inequality Lab.
Most people read this and think "the rich are getting richer."
But the strange question is how the rest of us ended up with so little?
Here's what I see:
In 2002, India's top 1% held 24.1% of the wealth and earned 16.7% of the income.
By 2022-23, it was 40.1% against 23.3%.
So what lifted wealth faster than income?
→ The researchers name the stock market boom as a likely contributor. The Sensex rose 7,300% between 1990 and 2023, and they say those gains are bound to reach mostly people with better access to the financial system. India had 1 billionaire in 1991. By 2022, it had 162.
→ Most household wealth sits in land and buildings. In the researchers' 2018 survey data, that was about 90%. Financial assets were about 7%, though surveys undercount them.
→ Professional portfolio management has come with a high entry ticket. SEBI's minimum for a PMS has been ₹50 lakh, and ₹1 crore for most AIFs.
→ Mutual funds did open a door from a few hundred rupees a month. In the 12 months to March 2026, the number of unique investors grew by nearly 78 lakh to 6.14 crore, and monthly SIPs crossed ₹32,000 crore (AMFI).
Getting in is only half the job. The other half is knowing what to hold, and how much, based on your whole financial life.
→ Account Aggregator is building the base for that. It lets people share bank, investment and insurance data in one consented view, and 5.96 crore people now use AA-powered finance management tools (Sahamati, FY26).
So far, this doesn’t show up in the wealth data. The World Inequality Report 2026 finds no sign of the gap closing in recent years.
I keep coming back to one gap: advice built around a family's whole financial picture, at a price a family can afford.
That is the gap we are building for at 1% Club. Find the link in the next tweet:
1
1
109
Since June 2022, loans taken by Indian families have grown 78%.
Savings grew only 49%. The whole economy grew 41%.
India's banks are the healthiest they have been in decades. The RBI is worried anyway. Here is why:
RBI's report has two numbers. Read the first and you relax. Read the second and you check your EMIs.
First number: 1.8%. That is bad loans at banks, meaning loans people have delayed repaying. It is the lowest in decades.
Second number: 58.4%. Of everything Indian families owe, 58.4% is personal loans, credit cards, gold loans and other retail loans that are not home loans.
A home loan has a house behind it. These depend more directly on your future income.
→ Loans taken for spending are nearly half of all family debt.
→ Gold loans are now the biggest part of it. Up 42.4% a year since March 2024, ₹5.14 lakh crore in May 2026. RBI says a long fall in gold prices could make repayment harder.
→ RBI has also flagged higher stress among borrowers who have taken loans from five or more lenders, with nothing pledged against them.
And savings are not keeping up. In FY26, Indian families saved 12.4% of GDP and borrowed 6.2%. That is about ₹50 borrowed for every ₹100 saved.
Nothing is breaking today. But loans with nothing pledged depend on your income, and gold loans depend on gold prices.
If your income stopped for three months, how many EMIs could your savings cover?
4
172
Eight months back, when we started FDE Academy, I had to explain what a forward deployed engineer is before anyone would listen to the rest.
This week India Today called it India's next engineer.
I think that tells you where this is heading. India built its tech name on engineers who sat in the back office and delivered to a spec. I see that changing. The engineer who sits with the client, learns how their business actually runs, and makes AI work inside it is the one companies are asking for now.
I don't think colleges or companies have figured out how to build that person yet. A curriculum moves in years, whereas this job moves in weeks.
So what does Indian engineering look like ten years from now?
Read the article here: indiatoday.in/jobs/in-focus/…
India just beat the US in PwC's AI survey by 19 points.
Every headline will call that a win. But it is the hardest news for a fresher.
I explained to News18 why:
PwC calls someone a front-runner if they have scarce skills and strong AI ability.
Globally, that is 14% of workers. In India, it is 2 in 5, the highest of any country in the survey.
So the person at the next desk is far more likely to be one. A fresher is judged against whoever sits closest.
Here is what that room looks like:
→ ICRIER surveyed IT firms. 55% say they have seen fewer entry-level jobs in the last two years.
→ LinkedIn's Grads' Guide 2026 says entry-level hiring at firms with 1 to 10 people grew 168% in India in two years. Indore, Jaipur and Bhopal are among the rising hiring hubs.
So some openings are shifting, from IT firms to very small firms and smaller cities. LinkedIn's advice to graduates is to show skills through projects.
If I were a fresher today, I would walk in with one thing
I built using AI. A small tool or workflow that saves someone real time. It gives an interviewer something concrete to react to.
The 40% shows where the pressure is, the 168% shows where the opening is.
1
8
343
Raghav Gupta retweeted
AIM Research and @FDE_Academy's latest report on India’s Forward Deployed Engineer talent market highlights a sharp hiring mismatch. Employer demand for field-deployed AI builders stands at 95%, while available talent is just 45%, creating a 50-percentage-point gap, the widest role-level shortage identified in the study.
The report says employers are struggling to find professionals who combine deep technical expertise with strong customer and stakeholder capabilities. It also notes continued retention pressure from remote opportunities, start-up premiums and entrepreneurial alternatives as organisations compete for production-ready AI deployment talent across India right now.
Read the full story here: analyticsindiamag.com/ai-hig…
#ArtificialIntelligence #AITalent #AIJobs #IndiaAI #TechJobs
Every wealth management startup in India is solving for the top 1% and calling it democratisation.
252 million people have linked their accounts to an account aggregator. UPI did 22 billion transactions last month. The actual middle class is right there in the data, and nobody is building for them seriously. This isn't because the market isn't big enough. It's only because it's harder and far less glamorous than landing HNI clients.
Honestly, that's one of the top reasons why we started 1% Club, and it's still the reason most of the competition doesn't scare me.
India is spending on AI faster than almost any country in the world and I think we're setting ourselves up for a very public reckoning in the next 12 to 18 months.
Enterprise AI investment grew 119% in one year here but only 22% of companies have any testing or risk assessment in place for what they're already running.
What I'm seeing is companies treating AI like a normal software rollout, buy the tool, plug it in, announce it, done. But these systems drift, and they fail without throwing an error. They make confident wrong calls that nobody catches because nobody built anything to catch them.
The ones moving fastest are usually the ones with the least guardrails and they think governance is something you sort out after you've won. It's not. Some of these deployments are going to fail very publicly and the boardroom conversation after that is going to be brutal for whoever signed off on it.
1
1
131
The biggest risk for the next generation of Indian engineers is not losing their job to AI. It's getting a really good AI job and having absolutely no idea what to do with the money.
I run a company that trains people for AI deployment roles and I'm watching this happen in real time.
Someone goes from earning 8 lakh to earning 35 or 40 lakh in under two years. Their parents never made that kind of money, nobody in their circle has, and overnight they're making financial decisions they have zero preparation for. New apartment, new car, lifestyle that scales with the salary, and three years later they have nothing to show for the best earning window of their life.
Everyone in the AI skilling conversation is obsessed with the job. How do you get the job, how do you keep the job, how do you move up. Nobody is asking what happens after. What do you actually do when you're 26 and your income has tripled and every financial instinct you grew up with was built for a completely different life.
That's why @getonepercent was built alongside @futurense. Both businesses look different from the outside but they exist for the same reason. Getting someone an AI-era job is step one, and ,making sure they actually build something lasting with the money from it is step two - almost nobody in India is connecting those two things together.
The job market and the money market are moving at the same speed right now and I think whoever builds for both at the same time has an unfair advantage over anyone doing just one.
1
94






