-
-
Assessment Question
Inspiration
The driving force behind SkillUp was identifying and addressing the Soft Skills Gap observed in students and young professionals. While these individuals are highly capable technically, crucial non-cognitive skills like collaborative communication, effective time management, and conflict resolution often lag, limiting their long-term career success. We recognized that personalized coaching for these essential skills is typically either generic or prohibitively expensive. Our goal was to leverage the power and accessibility of the Gemini AI to democratize professional coaching, creating a free, immediate, and highly personalized tool that translates self-reflection into actionable career intelligence, thus maximizing the project's Impact.
What it does
SkillUp provides an instant, structured, and personalized analysis of a user's soft skill profile. The core function involves the user completing a brief, 10-question self-assessment across five key skill domains. The system then routes the qualitative user responses through our Python Flask backend to the Gemini API. The AI, operating under a designated coaching persona, generates a detailed, customized report. This report is instantly rendered on the frontend in a clean, professional format, structured into four actionable categories: Identified Strengths, Areas for Growth, a Single Actionable Tip, and Real-World Resources for Improvement powered by Google Search Grounding.
How we built it
The application was constructed with a clean, decoupled architecture optimized for both Technical Craft and Feasibility. The frontend uses responsive design via HTML and Tailwind CSS, with JavaScript and the marked.js library handling the dynamic display of the AI-generated report. The backend utilizes a minimal Python Flask API to manage routing and the critical communication with the Gemini service. Key to the system's accuracy is the strategic integration of the gemini-2.5-flash model alongside a highly precise System Instruction that enforces the coaching persona and mandates a specific, four-part markdown output structure, ensuring reliable and high-quality advice.
Challenges we ran into
The primary technical challenge encountered was maintaining output consistency from the generative AI, which is essential for a predictable User Experience. We had to overcome the model's natural variability by rigorously refining the System Instruction prompt, essentially treating the prompt itself as code to enforce strict adherence to the four required output headers. Additionally, we addressed the potential for API instability, a common issue in cloud-based applications, by implementing exponential backoff and retry logic within the Python backend. This approach ensures the system remains resilient against temporary network latency or rate limits, significantly boosting the project's overall reliability.
Accomplishments that we're proud of
We are most proud of successfully integrating Search Grounding into the Gemini API call, a feature that profoundly enhances the project's Impact. By enabling the AI to access the live web, our tool is able to generate current, relevant external links (such as specific articles or videos) that directly address the user's identified growth area. This approach transforms generic feedback into a practical, actionable development plan. Furthermore, we take pride in the sophisticated User Experience, specifically the seamless assessment flow and the clean, custom-styled presentation of the complex AI-generated report, which was a significant Technical Craft achievement in frontend rendering.
What we learned
The project served as a powerful lesson in advanced prompt engineering, highlighting the difference between generating simple text and generating a structured, predictable data payload for an application. We learned that the AI’s System Instruction must be treated as a definitive technical contract, explicitly governing the required output format and maintaining the coaching persona. This knowledge about prompt constraints and the implementation of essential API resilience techniques like exponential backoff are crucial for developing production-ready, cloud-based tools, thereby solidifying the Feasibility of future SkillUp development.
What's next for Skill Up: The AI Soft Skills Coach
To ensure the project's long-term Sustainability and Impact, we plan to evolve SkillUp into a continuous soft skill development platform. Our immediate next steps include implementing user authentication to allow individuals to save and track their assessment results over time, providing a tangible metric for skill growth. Following this, we will introduce an Interactive Practice Mode utilizing Gemini's conversational capabilities to simulate real-world professional scenarios, such as practicing difficult negotiation or giving constructive feedback. Finally, we will develop modular, themed assessments to offer more specialized coaching in areas like Interview Preparedness and Advanced Leadership.
Log in or sign up for Devpost to join the conversation.