FDE_Academy

One of India's First Structured Pathway to Become an Forward Deployed Engineer
Built by real FDEs
PGP in FDE & Applied AI Solutions
Applications Open!!

[FDE_Academy @FDE_Academy]

Klarna’s AI handled two-thirds of customer chats in month one.

Then Klarna went back to recruiting human support agents.

Everyone is calling this an AI failure.

The AI passed the test, but the test might have been wrong.

Read to learn why:

In 2024, Klarna said its chatbot was doing the work of 700 agents, across 2.3 million chats in a month. Its CEO later said the bots were cheaper, but the quality was lower.

Why? One early tester said the bot mostly pulled answers from help documents and passed people to a human. Instant reply, customer still stuck.

Green on the dashboard. Stuck on the other end.

It’s not just Klarna. McKinsey’s 2025 survey found 88% of respondents use AI in at least one function, but only 7% say it’s fully scaled. The top performers were more than 3x more likely to aim for transformative change, not just efficiency.

A strong first month does not tell you how customers feel a year later. Klarna’s numbers and its CEO’s words point in different directions, and both are on the record.

[FDE_Academy @FDE_Academy]

TCS announced about 12,000 layoffs last year.

This July, it said it was building a team of up to 8,900 forward-deployed engineers.

Infosys wants 6,000 for a similar job.

Many say AI is killing IT jobs. Here is what they are missing:

One way to read it: the job is splitting in two.

→ TCS said its layoffs, at mid and senior levels, were about a skill mismatch.
→ The team-building plans are for the forward-deployed engineer (Infosys says "frontier engineers"), who sits inside a client's company and helps AI work in its real systems. One OpenAI New York posting lists discovery first, then design and build.

One reason the role exists: AI often stalls at deployment. In McKinsey's 2026 survey, 80% of respondents say AI improved their productivity. But only 6% met its ‘high performer’ definition: at least 5% EBIT impact attributed to AI, plus significant value.

TCS's CEO told Reuters that fitting AI into a large enterprise needs deep knowledge of client systems, which models don't provide on their own. He has a stake in that argument.

Scale check: TCS's target is 1% to 1.5% of staff, and Infosys's 6,000 is under 2%. Both are targets, not hires. The Reuters report has not linked TCS’s 12,000 layoffs to its up to 8,900 FDEs.

The job is about finding where a client's work gets stuck, then building the fix.

Hire or retrain: which one do you think TCS will pick?

[FDE_Academy @FDE_Academy]

https://t.co/BJ017yube0

[FDE_Academy @FDE_Academy]

San Francisco builds the AI.
New York hires the people who make it work.

Reportedly, 35% of FDE postings are in New York. San Francisco has 11%.

Here's what that tells you about the Indian situation:

An FDE works closely with customers to turn AI tools into functioning production systems. Per 2026 hiring guides, New York leads because of its banks and other regulated industries. A bank cannot afford a failed rollout, so it pays for someone to be there.

India's banks are in the same position, and the regulator is already involved:

→ In August 2025, an RBI committee released its FREE-AI framework, with 26 recommendations on how India's financial sector should adopt AI.

→ In June 2026, RBI released draft model risk rules. They say all models, including third-party ones, should be independently validated by the institution using them.

So a bank cannot simply plug in a vendor's AI tool. Someone has to make it work inside the bank. The bank must also check it independently and keep oversight. FDEs can help with the first part. The second part stays with the bank.

Indian companies are already hiring for it:

→ TCS plans up to 8,900 FDEs, around 1% to 1.5% of its staff.
→ Wipro plans 1,500+.
→ FDE and AI engineering jobs at India's global company offices are up roughly 1,000% in 12 months, per ANSR.

The best jobs will go to people who take that responsibility. Adding more headcount will not fill them.

[Raghav Gupta @Raghavv101]

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: https://t.co/afQ3dl5ftg

[FDE_Academy @FDE_Academy]

Giving every employee AI is like giving everyone a Ferrari and sending them through Silk Board.

Everyone gets a faster car. The junction stays the same.
That's why 8 in 10 people feel more productive, but only 6% of companies are high performers (McKinsey, 2026):

Two things show up in the same survey.

→ At an individual level, it works. 80% of respondents say AI improved their productivity, and an NBER study of 5,172 support agents found 15% more issues resolved per hour.

→ At the enterprise level, most respondents still don’t report a positive EBIT impact, 37% attribute any profit impact to AI, flat for a year, while the share scaling AI rose from 38% to 44%.

One reading is that the gap sits in the handoffs. Dessy Pavlova (TNW) describes a common setup: a customer arrives on the website, someone types it into a spreadsheet, another moves it to a second platform, and finance reconciles later.

AI can speed up several steps, but the handoffs stay.
McKinsey’s data is consistent with that reading: nearly 3 in 4 of the 6% say they fundamentally redesigned their workflows, against about 1 in 4 of the rest (survey data, not proof).

Knowing how to use AI can make one person faster.
Redesigning workflows was one thing most AI high performers had in common, and I expect hiring to move there: TCS plans 5,900 to 8,900 forward-deployed engineers, Infosys 6,000 (targets, not hires).

[FDE_Academy @FDE_Academy]

58% of GCCs in India take more than 45 days to fill critical AI roles and 50% are still making key hiring decisions without any data or predictive analytics. These are the same companies calling themselves AI transformation leaders for 2026.

The hiring infrastructure for this role hasn't caught up to the ambition and the cost of that gap is showing up in every delayed deployment, every failed rollout, every AI pilot that never makes it to production.

Building the technology turned out to be straightforward. Finding people who can actually execute on it is where companies keep getting stuck because they're still using old dev hiring playbooks.

[FDE_Academy @FDE_Academy]

58% of GCCs in India take more than 45 days to fill critical AI roles and 50% are still making key hiring decisions without any data or predictive analytics. These are the same companies calling themselves AI transformation leaders for 2026.

The hiring infrastructure for this role hasn't caught up to the ambition and the cost of that gap is showing up in every delayed deployment, every failed rollout, every AI pilot that never makes it to production.

Building the technology turned out to be straightforward. Finding people who can actually execute on it is where companies keep getting stuck because they're still using old dev hiring playbooks.

[FDE_Academy @FDE_Academy]

Accenture and Google Cloud just launched a joint business group and put 1,000 forward deployed engineers at the centre of it. Google is training every single one of them personally.

This is not a staffing decision. When two companies of this size build an entire business unit around one job title, they're telling you where the value in AI delivery has moved. It's not in the model. It's not in the infrastructure. It's in the person sitting inside the client's environment making it work.

[FDE_Academy @FDE_Academy]

https://t.co/f86PINNj7n

[FDE_Academy @FDE_Academy]

GPT-6 Astra dropped on September 3rd, and OpenAI is calling it the most intelligent and aligned model in the world.

72% of enterprises already have at least one AI workload in production. Only 6% qualify as actual high performers. The data suggests the bottleneck was never the model. Every major release makes the deployment gap more and more visible.

[FDE_Academy @FDE_Academy]

When the biggest technology and consulting companies are all competing for the same small pool of people, what does that do to the value of being one of those people?

[FDE_Academy @FDE_Academy]

Google earlier this year committed $750 million to embed its own FDEs across Capgemini, Cognizant and Deloitte. Now it's doing the same with Accenture. 1,000 more FDEs, trained directly by Google Cloud.

[FDE_Academy @FDE_Academy]

Every major consulting firm in the world is being restructured around this role right now. The question for engineers watching this is a simple one.

[FDE_Academy @FDE_Academy]

There were a lot of numbers in this report. But behind them is a bigger shift in how companies think about FDE talent.

We partnered with AIM Research to study India’s emerging FDE landscape demand, skills and talent gaps.

Read the full FDE Jobs Report: https://t.co/A8gIbXf1SL

[AIM @Analyticsindiam]

AIM Research and @FDE_Academy have found that 95% of Indian employers are looking for forward-deployed engineers (FDEs), the professionals who take AI systems from demo to deployment, but only 45% of the talent pool has these skills. That leaves a 50-percentage-point shortfall, the widest among all FDE role families studied.

According to the report, field-deployed AI builders command the highest median compensation among all experience levels, with pay rising from ₹32 lakh to ₹155 lakh. Bengaluru holds 34% of India's FDE talent, more than Delhi NCR, Hyderabad, and Pune combined.

Read the full report here: https://t.co/tP9dkoOhO4

#AI #India #TechJobs #ArtificialIntelligence #FutureOfWork #FDE #ForwardDeployedEngineer

[Raghav Gupta @Raghavv101]

We just released our first research report on the Forward Deployed Engineer talent landscape in India, done in partnership with AIM Research.

Here's what the data showed us: https://t.co/ndYqI1Qijp

Bengaluru holds 34% of all FDE talent in the country. More than Delhi NCR, Hyderabad and Pune combined.

70% of FDE professionals in India have 8 years of experience or less. Only 6% have more than 15 years. There is almost no senior bench anywhere in the market right now. Which means the engineers in that 4 to 8 year window right now are essentially the future leaders of this entire space. That window is open for maybe two or three more years before it gets crowded.

The biggest capability gap is not technical. Customer facing and consulting skills have an 88% employer demand against 55% supply. A 33 point gap. Core engineering has a 5 point gap. Everyone trained engineers to build, and nobody trained them to own.

Field deployed AI builders start at 32 LPA and go to 155 LPA at the senior level. Product companies pay 135 LPA at senior versus Big Tech's 85. The premium is on owning outcomes, not writing code.

68% of enterprise tech companies are already hiring for this role. So is 60% of BFSI. But healthcare is at 42%, manufacturing at 32%, retail at 36%. I didn't expect the non-tech sectors to be this far along already. This stopped being a software industry story a while ago and most people are still treating it like one.

95% employer demand for field deployed AI builders against 45% supply. A 50 point gap. The most constrained role in the entire ecosystem and the gap is projected to keep widening through 2027-28.

India has more engineers than almost any country in the world. The data shows we have a shortage of exactly one thing, people who can take an AI system out of a demo and make it work inside a real business. I'd read that shortage as the single biggest career opportunity in Indian tech right now.

[FDE_Academy @FDE_Academy]

The companies doing this are deploying faster. The ones still waiting are just falling further behind every month. If you want to be on the right side of that, FDE Academy is where we're training people for exactly this kind of work.

[FDE_Academy @FDE_Academy]

The early companies that went all in on the FDE model are starting to report actual results. An FDE at Superhuman shipped around 20 customer projects in four months. Deployments that used to take months are now taking weeks. Infor and Leaping AI are reporting similar speed gains

[FDE_Academy @FDE_Academy]

This matters more than all the hiring stats and salary numbers. The question was always whether embedding an engineer inside a customer's team would actually be faster than the old model of handing off software and hoping the customer figures it out.

[FDE_Academy @FDE_Academy]

This is the gap FDE Academy was built to close. If you're an engineer looking at this market and thinking about where the demand is actually headed, the data is already telling you.

[FDE_Academy @FDE_Academy]

Indeed's August 2026 data: FDE job postings are up over 5,000% compared to last year. Over 3,000 open roles on Indeed alone, and that's only counting postings that actually use "forward deployed engineer" in the title.

[FDE_Academy @FDE_Academy]

Meanwhile, a study shared with TechCrunch surveyed 250+ C-suite hiring executives across 180 companies and estimated the total global supply of qualified FDEs at roughly 2,000 people. Total. The demand is growing at 5,000% and the supply hasn't kept pace at all.