Inspiration
The realization that GitHub activity doesn't equal job-ready skill. Junior developers are swimming in unstructured busywork because they lack an objective, data-driven roadmap to bridge the gap between theory and production-level code quality.
What it does
Devbrah analyzes a developer's GitHub activity for code quality and skill signals. It then pinpoints competency gaps relative to the user's career goal and delivers a personalized, data-driven learning plan via an interactive dashboard and weekly digest, including direct links to necessary courses.
How we built it
Devbrah was engineered as a modern, full-stack application centered around a sophisticated code analysis pipeline and supported by cloud-native services. Our approach prioritized data-driven insights and a seamless developer experience.
Challenges we ran into
The primary challenge was filtering signal from noise in student repos—determining which commits actually reflected learning versus boilerplate. We also faced the dynamic curriculum problem, solving it by building an automated system to validate and tag external course resources against our evolving skill matrix.
Accomplishments that we're proud of
Successfully architecting and integrating the core pipeline: achieving a quantifiable metric for progress—the percentage of skills remaining to hit a stated goal. We built and validated the AI/LLM integration that powers the critique and recommendation system, all functioning end-to-end with our team's internal testing.
What we learned
Building Devbrah during the hackathon taught us that rapid technical integration is possible when components are clearly defined. We learned that mapping raw code metrics (like code complexity) to high-level skill inference requires careful tuning of the LLM prompts. Furthermore, we confirmed that the biggest technical hurdle is ensuring the AI's suggested curriculum is timely and relevant via reliable external data sourcing.
What's next for Devbrah
Next, we are launching Team Integration (B2B model) for engineering managers to quickly onboard and upskill new hires/interns against company standards as well as colleges to be able to upskill their students. We aim to secure $100,000 to finalize the B2B feature set and onboard our first 10 corporate clients.
Built With
- azure-openai-api
- fastapi
- get.tech
- github-oauth
- linkedin-oauth
- pytho
- react
- sqlite
- typescript
- vultr


Log in or sign up for Devpost to join the conversation.