Skip to content
L

LearnWithYuki

Yuki: Guiding Every Step Forward

Created on 20th September 2025

L

LearnWithYuki

Yuki: Guiding Every Step Forward

The problem LearnWithYuki solves

The Problem It Solves

Students often face challenges such as:

Information overload: Too many scattered resources without structure.

Lack of personalization: Generic roadmaps that don’t fit individual goals.

Limited guidance: Human teachers and mentors can’t scale to everyone’s needs.

Low motivation & consistency: No adaptive tracking or reminders to keep learners on track.

What People Can Use It For

Personalized Study Roadmaps: Automatically generated plans based on goals (e.g., exams, skills, or self-learning).

AI Subject Teachers: On-demand explanations, Q&A, and tutoring support.

Progress Tracking: Milestones, reminders, and adaptive scheduling to improve consistency.

Goal Achievement: Helps students break down long-term objectives into actionable daily/weekly tasks.

Tiered Features: Free users get structured plans, while premium users unlock deep analytics, interactive tutors, and advanced strategies.

How It Makes Tasks Easier & Safer

Efficiency: Saves time by generating structured learning paths instantly.

Clarity: Reduces confusion from random internet resources by offering focused guidance.

Adaptability: Plans evolve with learner progress, ensuring continuous alignment with goals.

Accessibility: Provides teacher-like mentorship without geographic or time restrictions.

Scalability: Safely supports thousands of learners with AI-driven automation.

Challenges we ran into

Multi-Agent Workflow Development:
One of the biggest hurdles we faced was building the agentic workflow. Initially, we experimented with CrewAI, but scaling the system proved difficult. We tried multiple SDKs, including OpenAI’s SDK, before finally settling on LangGraph, which provided the flexibility and scalability we needed for multi-agent orchestration.

Context Management for Users:
Another challenge has been handling the context size of user interactions. Since LLMs have limitations on input length, ensuring that user history and goals remain relevant without exceeding limits is tricky. To overcome this, we are working on query optimization, response summarization, and efficient context window management to keep interactions smooth and scalable.

Tracks Applied (1)

$300(Open): Cash Prize

Startup Mindset: We are not just building a project, we are laying the foundation for a future startup. The hackathon gi...Read More

Discussion

Builders also viewed

See more projects on Devfolio