A real story
Aanya graduated from a Tier-3 engineering college in Indore with a 7.4 CGPA. Through the entire placement season, four companies visited her campus. Two filtered everyone below 8.0 CGPA in the first ten minutes. The other two ran coding tests that felt nothing like the textbook DSA she'd crammed. By December, she had zero offers and a WhatsApp group full of equally anxious batchmates.
The story she told herself: there are no jobs. The data tells a different — and ultimately more hopeful — story.
56.35%
National employability 2026, up from 46% in 2022 (India Skills Report)
42,000+
Freshers TCS targets hiring in FY26 alone
600%
Growth in AI/ML job listings (India Skills Report 2026)
60–70%
Of 2026 fresher JDs list AI/ML familiarity as baseline
Employability is rising, not collapsing
The India Skills Report 2026 puts national employability at 56.35%, up from 54.81% last year and ~46% in 2022. The median graduate is slowly becoming more job-ready.
The gap is brutal at the bottom
Mercer-Mettl's India Graduate Skill Index measured employability as low as 42.6% in 2024, with most 'employable' candidates coming from Tier-1 cities. Large share of graduates simply cannot demonstrate what employers test for.
The screen has changed
Recruiters are quietly retiring CGPA filters (7.0+, 7.5+, 8.0+) because CGPA poorly predicts performance and manual screening doesn't scale. In their place: practical coding assessments, GitHub/Kaggle portfolios, and aptitude tests.
More total doors, not fewer
Campus drives, TCS NQT, off-campus rounds, skills-based platforms, gig and project work. The freelance workforce is projected to hit 23.5 million by 2030, with project-based hiring up 38% in a year.
The honest summary
The doors don't open for a CGPA. They open for evidence.
"Learn a new skill" is useless advice without specifics. Here's where demand and accessibility actually overlap for a fresher starting today:
AI / ML literacy & prompt engineering
Every engineering and analytics fresherBaseline in 60–70% of fresher JDs; AI/ML roles up ~600%
Data analytics (SQL, Python, dashboards)
Commerce, science, engineering gradsBFSI, fintech, healthcare analytics all expanding
Cloud & DevOps (AWS/Azure/GCP, Terraform, CI/CD)
CS, IT, ECE gradsPersistent shortage; clean portfolio path
Cybersecurity
CS/IT — anyone willing to study itNamed a top in-demand skill in 2026 reports
Performance marketing & SEO
Commerce, arts, any streamCommerce employability jumped to 62.81% partly because of this
Commerce employability jumped to 62.81%, science to ~61%, and arts to 55.55% precisely because digital and interdisciplinary skills opened roles that a degree alone never would. You do not need to be a CS topper. You need one demonstrable skill stack and the proof to back it.
Here is the unfair advantage almost no fresher in a Tier-2 or Tier-3 town is using: you can run a genuinely capable AI model on the laptop you already own, for free, offline, forever. In 2026 the open-weight model ecosystem matured to the point where local models handle 70–80% of everyday coding and analysis tasks at zero API cost.
8–16 GB RAM laptop
40–60 tokens/sec on modest laptops — fully usable
16 GB + general purpose / vision
Strong for reasoning and multimodal tasks
Good GPU (RTX 4090) or M-series Mac
Near-commercial quality, zero API cost
Indian language workloads
Strong interview talking point for India-focused roles
Licence note
Apache 2.0 and MIT models (Qwen, Phi, GPT-OSS, Sarvam) are the cleanest to build on and even commercialise. Gemma uses a custom licence — fine for learning and internal tools, but read the terms before shipping a product.
Install Ollama (ollama.com)
The simplest path. One command pulls and runs a model: ollama run qwen3-coder or ollama run phi4-mini. Prefer a GUI? Use LM Studio. Want maximum speed? llama.cpp.
Talk to it, then build with it
Ask it to explain code, refactor a function, write tests, or draft SQL. You're now learning the exact 'AI-assisted development' workflow employers expect.
Build something small and real
A resume-screener, a study-notes summariser, a customer-support bot for a fictional shop, a RAG system over your college's PDF notes. Each one is a project.
A recruiter spends roughly six seconds on a fresher's résumé. In those six seconds they are not reading your CGPA — they are looking for clickable links to things that work. Proof of work is verifiable, public evidence that you can build, not just describe.
Software / AI
GitHub repos + a live deployed app (Vercel or Netlify, free in 60 seconds)
Data / analytics
Dashboards on Tableau Public or a Kaggle profile with notebooks
Cloud
A GitHub folder of real AWS/Azure deployments with architecture diagrams and clear READMEs
Design
Behance or Dribbble — never a Google Drive of PNGs
Marketing / growth
A Medium or Notion writeup of a real campaign, an SEO experiment, a side project's analytics
Pick projects aligned to your target role
Applying for frontend? Lead with UI projects. Data analyst? Lead with dashboards. Recruiters skim for relevance.
Pin 3–5 best repos on your GitHub profile
That's your highlight reel — the first thing a recruiter sees when they click your link.
Write a real README for each project
What problem it solves, the tech used, your role, and a live demo link. A project without a README is a recipe with no instructions.
Deploy the top two live
A clickable URL the recruiter can poke on their phone beats ten lines of 'skills' on a résumé.
Write like a human, not a robot
'I built this because I was tired of tracking expenses in Excel, so I made a tracker with login and reminders' beats 'Built a web app using React.' People hire people.
Put the links everywhere
Résumé (next to each project), LinkedIn Featured section, email signature.
A useful reframe
A degree says what you studied. Proof of work says what you can do. In 2026, with CGPA filters fading and skills-based screening rising, the second one wins interviews.
A concrete sequence beats vague ambition. Adapt the role, keep the structure. India's peak hiring windows are January–March and September–November — time this plan accordingly.
Aanya didn't get hired in December. She got hired in February — by a Pune SaaS startup that found her through a small, well-documented AI project she'd built on her own laptop and shared on LinkedIn. The hiring manager told her something she repeats to every junior now:
"Your CGPA got you ignored. Your repo got you the call."
Campus offers feel fewer because companies shifted from bulk hiring to selective, skills-based hiring. But total hiring across IT, BFSI, manufacturing, pharma, and the gig economy is healthy. TCS alone targets ~42,000 freshers in FY26. The scarcity is in job-ready candidates, not jobs.
No. You can download an open-weight model (Phi-4-mini, Qwen3-Coder, Gemma 4, GPT-OSS, or Sarvam for Indian languages) and run it free on a 16 GB laptop using Ollama. Learning by building costs ₹0 and produces proof of work at the same time.
It's a powerful signal, not a magic bullet. Pair a strong, well-documented portfolio with a clean résumé and a focused skill stack. Most recruiters treat it as the evidence that turns a 'maybe' into an interview.
Increasingly, no. Many 2026 employers have dropped CGPA cut-offs in favour of coding assessments and project portfolios, because CGPA poorly predicts performance. Proof of work is how you route around the filter.
Performance marketing, SEO, CRM, and growth operations are expanding fast, and commerce graduates' employability rose to 62.81% on the back of BFSI and fintech hiring. Data analytics is valuable across almost every stream.
Find out how AI will affect your target role
Take the free AI Risk Quiz — understand your risk score and which skills to build first.
Take the Free AI Risk Quiz →