Inspiration
Saksham AI was inspired by the enormous skill and career guidance gap faced by youth in India’s rural and semi-urban regions. With millions of graduates lacking direction, especially in Tier 2 and Tier 3 cities, and a large percentage remaining unemployable due to insufficient access to high-quality learning and placement support, the founders of Saksham AI decided to build an agent that empowers students with actionable, personalized skilling and career recommendations using AI and public data resources.
What it does
Saksham AI acts as a multilingual, conversational career and skilling guidance agent. It: Recommends tailored learning pathways and soft/hard skill courses (from SWAYAM, NSDC, etc.) based on user profiles. Suggests localized job opportunities by analysing hiring trends and demand. Answers queries on government skilling schemes like PMKVY, Skill India, etc. Supports natural language interaction in multiple Indian languages, using translation models for region-specific accessibility. Provides stepwise, context-aware guidance rather than static information.
How we built it
Leveraged IBM’s Agent Development Kit for stateful, context-aware dialogue workflows. Integrated public data from national job boards, NSDC, SWAYAM, and others to power recommendations. Incorporated LLMs (like Granite) and IndicTrans2 for robust natural language support across Indian languages. Combined Socratic tutoring models for personalized, adaptive skill-building. Architected an open-source platform for scalability and community-driven enhancements.
Challenges we ran into
Ensuring the platform offers dynamic, real-world oriented exercises was challenging; some users found initial problem sets too simple or repetitive. Technical difficulties included slow platform response times, bugs, and UI inconsistencies (like losing progress on refresh). Making AI guidance genuinely actionable and detailed, especially for diverse user backgrounds. Building trust in AI-powered advice and ensuring accurate translations across many Indian languages. Scaling the solution to represent the full diversity of Indian educational backgrounds and needs.
Accomplishments that we're proud of
Successfully piloted AI-driven personalized skilling journeys for students outside Tier 1 cities, closing the employability gap. Developed a multilingual agent capable of meaningful natural language interaction for first-generation learners. Aggregated and simplified access to government skilling schemes, previously scattered across multiple portals. Fostered active engagement and feedback loops with real users (over 1,000+ participants in pilot studies). Contributed to the goal of reducing India’s underemployment rate and aligning student skills with actual job market needs.
What we learned
Personalized, context-aware AI can dramatically improve engagement and outcomes for underserved youth, but must be continually refined based on user feedback. Language inclusivity is critical – accessible, mother-tongue guidance removes huge barriers to adoption. Simple interfaces, community features, and actionable, stepwise recommendations are effective for non-expert users. Scaling requires not only technical excellence but also cultural/contextual adaptability and ongoing research partnerships.
What's next for Saksham AI
Broaden integration with additional local and regional employment data sources for even more precise recommendations. Enhance adaptive learning features: richer feedback, community-based peer learning, and gamification. Expand curriculum breadth and real-world challenge variety. Strengthen technical performance, add more languages, and deepen support for government upskilling. Pursue institutional partnerships for mainstream adoption and impact measurement across India and beyond.
Log in or sign up for Devpost to join the conversation.