CurrAlign AI is an on-device, explainable AI system that evaluates how well an academic curriculum aligns with real-world industry job requirements.
It compares syllabi with job descriptions, identifies skill overlap, missing competencies, and low-relevance topics, and explains why these gaps exist — all without cloud dependency.
Academic curricula evolve slowly, while industry skill demands change rapidly.
Institutions lack a simple, objective, and transparent way to measure curriculum relevance, and students have limited visibility into employability gaps.
CurrAlign AI provides:
- Automated skill extraction from syllabi and job descriptions
- Curriculum–industry alignment analysis
- Explainable reasoning for every insight
- Qualitative readiness assessment for decision-making
All analysis runs locally on the user’s device.
- ✅ Explainable AI — no black-box scoring
- ✅ On-device execution — privacy-first, low latency
- ✅ Deterministic pipeline — predictable, auditable results
- ✅ Education-focused — built for institutions, not hiring platforms
- ✅ AMD-aligned architecture — optimized for local CPU/GPU compute
- Upload syllabus and job descriptions
- Extract and normalize text
- Identify skills and topics
- Compare curriculum vs industry demand
- Generate readiness summary and explanations
- Display on-device performance characteristics
- Frontend: React, Tailwind CSS
- Backend: FastAPI (Python)
- NLP: TF-IDF, keyword extraction, cosine similarity
- Document Processing: pdfplumber, python-docx
- Execution: Fully on-device (no cloud services)
- Designed for efficient execution on AMD multi-core CPUs
- Parallelizable NLP and document processing workloads
- Privacy-preserving edge AI architecture
- Ready for future AMD GPU acceleration
Includes a One-Click Judge Mode with preloaded demo data for instant evaluation.
🎥 Deployment : [(https://adarsh-gautam-sys.github.io/CurrAlign-AI/)]
🔗 GitHub Repository: [(https://github.com/adarsh-gautam-sys/CurrAlign-AI)]
CurrAlign AI transforms curriculum evaluation from a manual assumption into a measurable, explainable process.
Built for the AMD Slingshot Hackathon 2026 🚀