LernexAI

LernexAI is a full-stack ed-tech OS pairing sub-500ms AI tutoring with proctored exams, real-time anti-cheat heuristics, and verifiable crypto credentials on an ultra-efficient, zero-burn $1 stack.

Comment

Inspiration

In emerging tech hubs like India, the conventional EdTech landscape is flooded with monotonous, multi-hour pre-recorded video lectures wrapped in intimidating English jargon. This systemic language barrier and cognitive "video fatigue" cause millions of aspiring developers from tier-2 and tier-3 cities to lose confidence and abandon computer science.

I wanted to dismantle this friction entirely. That is why I built LernexAI—an interactive, hands-on educational operating system designed to deliver world-class technical education in crystal-clear, approachable native languages (like Hinglish) and English. No passive 40-minute video marathons, no jargon anxiety—just active typing, real-time problem solving, and deep engineering mastery in the language students actually think in.

What it does

LernexAI shifts the learning paradigm from passive observation to active engineering execution. The entire ecosystem is zero-setup and operates seamlessly in the browser:

  • Interactive Multi-Language Sandboxes: Students write, compile, and debug real code in C, Python 3.12, Java, JavaScript, and Enterprise SQL directly in the browser with real-time hardware execution logs.
  • Sub-500ms Context-Aware AI Tutor: The AI reads the student's active lesson context, code syntax, and runtime errors in real-time. It delivers bilingual hints and pedagogical breakdowns without spoon-feeding exam answers.
  • Proctored 2-Strike Assessment Engine: A robust anti-cheat testing room that tracks focus loss, detects unauthorized tab-switching, and embeds dynamic candidate forensic watermarks. Strike 1 issues an urgent warning; Strike 2 automatically submits the test.
  • Cryptographic & Verifiable Credentials: Graduates earn certificates designed with classical anti-counterfeit Guilloche rosettes, unique verification hashes, and public QR endpoints verifiable by employers.
  • Automated Zero-Cost Helpdesk: An intelligent support desk that runs real-time NLP classification on student tickets. Admins reply via a native "Swipe-to-Reply" Telegram integration that syncs back to the student's web dashboard instantly.

How we built it

LernexAI is engineered with an enterprise-grade, zero-leak architectural philosophy on an astonishingly lean budget:

  • Frontend: React 18, TypeScript 5.5, and Vite, styled using Tailwind CSS and Framer Motion.
  • Backend: Dual-compatibility architecture leveraging Vercel Serverless Functions and Express.
  • Database & Security: Supabase (PostgreSQL) with strict cryptographic JWT authentication and hardened Row-Level Security (RLS) policies.
  • Resilient AI Matrix: Groq Cloud SDK (Qwen & Llama) and Google Gemini Flash powered by a custom Atomic Round-Robin Pooling architecture to ensure zero-downtime inference at zero cost.

Challenges we ran into

  • The $1 Production Budget Ceiling: Deploying a platform with real-time AI streaming, proctored testing, and database isolation without exceeding a $1 domain cost required relentless resource optimization.
  • Latency & Reasoning Purification: Modern reasoning models output bulky chain-of-thought blocks. We engineered an asynchronous pipeline to sanitize thinking tokens on the fly, keeping our AI Tutor's latency under 500ms.
  • Proctoring Heuristics in Browser: Preventing unauthorized DevTools and multi-tab switching while avoiding false positives was tricky. We carefully calibrated debounce thresholds for focus blur events across different browser engines.

Accomplishments that we're proud of

  • True Zero-Cost Infrastructure: Successfully architected and deployed a highly complex, real-time EdTech framework under a strict $1 budget.
  • High-Speed AI Pedagogy: Achieved consistent sub-500ms tutor response times using our multi-key round-robin Groq LPU engine.
  • Legitimate Academic Credentialing: Built dynamic vector certificates with instant public QR verification.
  • Production Rigor: Maintained strict code health and version control through 78+ granular git commits and zero TypeScript compiler warnings.

What we learned

  • High-Throughput Key Pooling: We mastered the engineering of atomic round-robin dispatch queues (currently powering our AI Tutor and Help Center) to maximize free-tier cloud quotas safely and ensure 100% API uptime.
  • Fine-Grained Database Security: Designing PostgreSQL RLS policies to prevent privilege-escalation attacks in multi-tenant environments.
  • Vernacular AI Pedagogy: Crafting strict system prompts that guide LLMs to speak naturally and encouragingly while maintaining collegiate-level technical accuracy.

What's next for LernexAI

  • The Ultimate "Tech-Only" Hub: Cultivating a hyper-focused, noise-free ecosystem strictly dedicated to computer science, coding, and software engineering—eliminating generic courses to build a high-quality community of pure tech talent.
  • AI-Powered Creator Studio: Launching a public course-building engine where developers, educators, and tech creators can auto-generate and publish their own interactive coding modules and proctored tests for the global community.
  • Grassroots Engineering Empowerment: Bringing hands-on computer science learning to underserved students in Tier-2/3 institutions who lack campus mentorship and heavy hardware resources.
  • Decentralized Verification (SBTs): Minting proctored assessment hashes directly to EVM chains to provide tamper-proof, globally recognized credentials for hiring.

Built With

Share this project:

Updates

Submission history