How Scaler is Navigating the AI Era of Free Upskilling
By Supreeth Koundinya ·
Even before the AI era, the internet and platforms like YouTube had already opened up access to education at scale. Now, generative AI tools such as ChatGPT have further widened that access. Learners can now ask questions and get explanations in real time.
So, where does that leave a structured edtech platform like Scaler?
In its latest overhaul, the edtech firm, valued at $710 million, is positioning itself as fully AI-native, rebuilding its programmes from the ground up rather than adding AI modules to existing courses. Even then, what does an AI-native edtech platform look like?
The Execution Gap in AI Upskilling
In a joint study with CyberMedia Research, which surveyed 400 software engineers and tech recruiters, Scaler found that candidates could explain concepts, list tools, and complete guided exercises, but struggled to build systems, debug failures, and make design decisions when requirements changed.
In an environment where India’s AI talent demand is projected to nearly double by 2027, according to Deloitte and NASSCOM, Scaler found that while there is plenty of access to upskilling, the gap exists in execution.
In an interaction with AIM, Amar Srivastava, Scaler’s CEO of Online Business, explains that while free resources and AI tools reduce the cost of learning, they do not enforce practice, evaluation, or feedback. “We not only offer content, but we offer content that is industry-ready and delivered by industry experts,” he says.
Scaler’s model is built around three layers: learning concepts, applying them repeatedly, and testing them in conditions that resemble real work. “It is only 30% what you need to learn from instructors. The rest 50% is practising day in and day out,” Srivastava notes.
He observes that a gap lies in when learners move from understanding concepts to implementing them in real scenarios, particularly when systems break or behave unpredictably.
To address this, Scaler sources assignments from real company problems and refreshes every time a class is run, so the same topic comes with updated problem sets. This is possible through partnerships with over 1,200 employers across India and globally, including large tech firms such as Meta, Amazon, and Apple, as well as startups and frontier AI labs working on advanced systems. It also gives access to Mentors and teaching assistants, who are working engineers, alongside peer groups moving through the same material.
Srivastava says the platform tracks how learners solve these and intervenes at specific points where they struggle. Personalisation is driven by historical learner data, adapting what each learner focuses on next.
From Scaler 1.0 to AI-Native
Srivastava describes the current shift as the third iteration of the platform.
The first phase focused on data structures and algorithms, aligning with a hiring environment where companies prioritised problem-solving and fundamentals. This led to strong hiring outcomes, including large intakes from companies like Amazon, where he claims that the company recruited more from Scaler than all IITs combined.
“But slowly, we realised that in 2023–24, the industry needed talent that not only understood fundamentals, but also system design, even at a junior level,” he remarks.
As hiring expectations evolved, companies began looking for engineers who could work across system design layers such as LLD (low-level design) and HLD (high-level design) and handle advanced areas in data science, rather than just fundamentals.
This shift led to Scaler’s second phase, where the curriculum expanded to include production-level skills and pushed learners to spend more time building depth.
The current iteration, Scaler 3.0, builds on that, with AI integrated into how these roles operate. Srivastava observed that while there is a perception that AI fluency alone is enough, companies still expect strong fundamentals, now applied within AI-driven environments.
Software engineers are now building systems that integrate LLMs into the development lifecycle. In data science, the focus has moved to working with transformer models, fine-tuning them, and deploying them in production environments. In DevOps, roles now include managing AI-driven infrastructure and platforms.
Srivastava says the shift shows up in how each role is taught. Interview preparation has changed to include AI-based coding assessments and system design problems where AI is part of the architecture.
The curriculum is also updated continuously based on hiring signals and ecosystem changes. “Every three months, we do a formal overhaul of the curriculum,” the executive notes.
Workshops, masterclasses, and elective modules are used to bring new frameworks and approaches into the system quickly, before they are absorbed into the core curriculum. For instance, when OpenClaw was released and gained traction, Srivastava says Scaler moved quickly to incorporate it through masterclasses and internal sessions, focusing not just on how it works but how learners can build similar systems.
This also extended through Scaler AI Labs, which works closely with such teams, giving learners exposure to ongoing work and opportunities to contribute to real-world problems, including those that are ahead of current hiring requirements.
Assignments follow the same pattern. Srivastava says that even when the same concept is taught again, the problems are updated to reflect what companies are currently testing for.
Outcomes, ROI, and Uncertain Future
However, when students invest ₹2.5–3 lakh in Scaler’s courses, they typically expect a clear return on investment. Most often in the form of securing employment.
Srivastava says the company structures its programmes over 12–18 months, arguing that shorter formats are not sufficient to build meaningful capability in the tech ecosystem.
It runs a dedicated placements team of 30–40 people who begin working with learners after they complete internal assessments. This includes resume building, interview preparation, and matching candidates with opportunities through its hiring network.
While the company does not guarantee placements, Srivastava cites internal data showing that 96% of learners who completed the programme in its last report secured career opportunities. He notes that demand from employers is increasingly skewed towards AI-native roles, with companies indicating that such talent is limited and often comes at a premium, which the platform is trying to address through its latest overhaul.
And that shows up in market estimates. According to a report by Quess Corp, the AI talent shortage in global capability centres has hit 42%.
However, at the same time, the assumption that more engineers need to be trained for an AI-driven future is not universally agreed upon. Some technologists predict that AI could automate large parts of software work within the next 12–18 months, raising questions about the long-term demand for traditional engineering roles. Others argue the shift is less about replacement and more about AI taking over tasks while increasing the need for engineers who can design systems, review outputs, and make decisions.
Either way, Scale not only has to one-up other edtech platforms delivering AI courses, but has to constantly reinvent its teaching methodologies to counter free AI education.