NorthStarz.ai Is Focused on Closing India’s Soft-Skill Gap Through AI-powered Screening

G Reshma

July 7, 2025

India’s education system continues to remain domain-centric while giving little structured attention to soft skills that corporates look for when hiring. For Avinash, a former HR leader with over two decades of experience, this disconnect was as clear as daylight in the hiring process, with candidates often arriving unprepared, lacking the traits companies valued most.

Avinash says, “Around 6.5 lakh people graduate every year, and 85 percent start looking for jobs. But when companies hire, they rarely find candidates who are prepared.”

This gap led to the NorthStarz.ai platform, owned by Novenue Technologies and founded by Saurabh Sisodia Singh, Avinash V Singh, and Rajiv Ranjan. It started off as an attempt to help students self-assess and improve their interview readiness, and later evolved into an enterprise product that automates early-stage candidate screening.

Based in Bengaluru, NorthStarz.ai now evaluates both technical and human-centric skills through AI-led interviews, helping companies identify fit with greater precision. The platform used by companies like Mahindra Logistics has processed over 60,000 candidate assessments, and has raised 250K in funding.

Shift from student upskilling to enterprise-focused screening

After jumping in and out of corporate roles and entrepreneurial stints, Avinash and his co-founders were set on building something that would reshape how companies worked, where the discussions kept circling back to the same problem. While capital was available in abundance, good human capital was still missing.

Avinash says, “The question eventually boiled down to why the failure rate of startups in India is so high, and it wasn't capital that was the problem anymore. What was missing was good human capital. Seven to ten years ago, the quality just wasn’t very innovative.”

Diving deeper into how companies could identify high-potential talent uncovered an old problem. The education system was built around domain knowledge, with little focus on the soft skills companies actually looked for. That gap led to the creation of Interview Vision, a platform designed to help students assess their soft skills and prepare for interviews.

Avinash explains, “Around 6.5 lakh people graduate every year, and 85 percent start looking for jobs. But when companies hire, they rarely find candidates who are prepared. With over 22 years in HR, we saw this problem repeatedly. So we thought, why not start there and build a platform for candidates to help them see where they stand?”

They started with college seminars in tier 2 and tier 3 cities, but the platform did not gain the traction they had expected, which made Avinash rethink the approach and explore what could be done differently. In that process, he arrived at the idea of serving corporate clients instead of students, which set things in motion for what NorthStarz.ai is today.

“When the platform did not work out with the students, we thought, why not flip the whole ecosystem? If candidates are not interested, let's go to companies because companies want to identify such skill sets in human beings when they're hiring.”

How NorthStarz AI turned soft skills into structured intelligence

NorthStarz.ai interviewed over 60,000 candidates to identify five core skills experts value most: critical thinking, growth mindset, active listening, communication, and teamwork. They arrived at these through conversations with psychologists, HR leaders, and long-time industry partners.

A parallel study of company value systems showed these five traits reflected what most organisations prioritised. These responses were scored, reviewed, and standardized. That data became the foundation for training the company’s AI models.

Avinash explains, “We also started researching on the kind of values that drive companies and what it is that most companies use amongst their cultural or you know, value system. And we found that while there are eight to 10 larger aspects, these five broadly capture almost 60 to 70% of every company's value system. And hence, we narrowed down to these five.”

NorthStarz.ai evaluated and standardized the expert scores, then fed that data into its own AI model, built specifically to identify the five core skills at a time when domain-specific AI was not even a known concept, as the team knew that general-purpose AI could not reliably measure traits like intelligence or soft skills.

“Once we did that, and it took a lot of time because we built our data, created our models, trained the system, we realised we had built something no company in the world had at the time”

AI-led screening system built for scale, flexibility, and recruiter relief

NorthStarz.ai allows recruiters to post jobs based on specific role requirements, after which candidates receive a unique interview link they can access at their convenience. The platform conducts the full interview, assessing both technical and human-centric skills aligned to the job description.

Built with compliance in mind, it follows India’s DPDPA and approved communication standards, including those for WhatsApp, ensuring a secure and flexible experience for candidates.

Avinash explains, “We did that by automating the entire communication flow on WhatsApp, and we eliminated the need for the recruiter to take the initial interview. This interview link can be, you know, used to take an interview anytime by the candidate, be it day, be it night, because a human is not conducting the interview, it is being conducted by the AI itself.”

NorthStarz.ai evaluates candidates across four key dimensions:

Resume fitment
The platform checks how closely a candidate’s resume aligns with the requirements of the job description.
Employability skills
It assesses human-centric traits such as communication, critical thinking, and adaptability based on how candidates respond in interviews.
Technical fitment
Candidates are tested on their knowledge of relevant tools, technologies, and domain-specific skills needed for the role.
Confidence analysis
Through voice analysis, the system evaluates how confidently a candidate answers each question, adding a layer to the assessment.

“The fourth is something we built entirely in-house, and no other platform offers it today,” he says. “We can assess a candidate’s confidence on each answer by analysing their voice, things like pitch, volume, and filler words. It helps us understand how sure they are about what they’re saying.”

Once the individual scores are generated, NorthStarz.ai aggregates them and ranks candidates on the recruiter’s dashboard, removing the need for manual screening. Avinash points out that candidates often tailor their resumes to match job descriptions, making it hard for recruiters to distinguish genuine fit from surface-level alignment. As a result, they end up calling far more candidates than necessary.

“NorthStarz.ai automates three key tasks that recruiters typically handle manually which rescreening resumes, evaluating candidate skills, and ranking applicants based on fit and since candidates use resume builders recruiters that had to call 20 people now tell us call 100 unsure who to shortlist and its reached a point where they say it feels less like hiring and more like working in a call center.”

NorthStarz AI defines its core success metric as the number of candidates who complete an interview on the platform. Everything else hinges on that outcome.

“Our North Star metric is the number of candidates who complete an interview on our platform. That’s the only thing which is sort of the success factor, not just for me, but for my customer as well.”

Sector focus and volume-linked pricing structure

NorthStarz.ai primarily works with companies across ITES, BFSI, GCCs, and staffing or augmentation firms. These sectors form the bulk of its current user base. Within ITES, typical roles include software engineers, data engineers, back-end and front-end developers. In BFSI, the focus shifts to sales roles, field agents, business development executives, and operations or back-office staff.

In the staffing, augmentation, and BPO segments, the platform is used to assess candidates for customer care, support, and general operations roles. These companies rely on NorthStarz.ai to screen for critical soft skills before hiring at scale, across high-volume functions.

Avinash says, “AI adoption matures, more profiles would be added to their systems. So, yeah, I mean, we are seeing an upwards tick in the number of companies who are now willing to come and adopt AI or at least evaluate it and find out how it can help them.”

NorthStarz AI prices its product using a freemium model tied to interview volume, and every company gets one job listing free each month. Beyond that, pricing is based solely on the number of completed AI assessments, and companies can purchase a monthly subscription starting from Rs. 13,200 to custom pricing packs with fixed interview counts, allowing costs to scale with usage while paying only for actual output.

Avinash says, “The product follows a pure freemium model. Anyone who visits our portal gets one free job listing per month. Beyond that, we charge based on the number of completed interviews. We only bill for completed assessments on our platform.”

From Python to Go (Golang) - Meeting performance demands

NorthStarz AI was built before generative AI became widely accessible, and with no reference architecture in sight, the team had to build everything from scratch, where every component of the assessment engine was shaped through trial and error.

With no existing datasets of interview responses to draw from, the team spent two years manually conducting and processing interviews to train their models and get them production-ready.

“We had to write our machine learning algorithms from scratch. This was something unique because nobody else had done it in the world. Nothing like this existed as a reference to us.”

The tech stack had to evolve owing to its non-scalable nature. Python, which was the language used earlier and the standard in data science, fell short on performance at scale. This made NorthStarz AI pivot its entire stack to Go (Golang), requiring a complete migration of the backend to a language better suited for speed and concurrency.

“Python has been shouted out as the language for data engineers, which is incorrect. We find fewer candidates who are experts in Go, so we often have to train them ourselves.”

Data also had to be created the hard way, manually conducting and processing interviews over two years to get their models production-ready.

Tracking the shift from general-purpose to specialised and autonomous AI

NorthStarz AI is closely tracking the shift towards domain-specific models where custom-built systems are trained on proprietary data rather than relying on generic large language models. That tailors intelligence for their internal workflows, it sees as foundational to the next phase of enterprise AI, where Avinash refers to domain-specific AI as the flavor for the near future. The next thing is the use of agentic AI.

“You see, companies build their own AI models on their internal data, which is fit for their use. For example, Accenture is building its own AI model. JP Morgan is building their own AI model. Apple is building its own AI model. These are primarily models that are being built to help these companies succeed by leveraging the data that these companies have generated within their lifespans.”

The organisation is also looking to leverage agentic AI, where autonomous agents act across internal tools and departments to surface risks, flag opportunities, and support faster decision-making. As part of this, it is building what it calls the CEO Command Center, a system of agentic bots built on retrieval augmented generation that integrates with each department’s tech stack.

“What happens is that the responsibility to generate output from multiple sources ends up on human beings. And human beings are not always the best when it comes to reporting. These agents can be taught how to identify flags, how to spot opportunities, and how to flag concerns, and they will do it in their rawest and sharpest form, which human beings cannot.”