Joblet

Inspiration

We were inspired by a simple problem: there is more information available than ever before, yet making clear decisions is still difficult. News, earnings reports, and market updates are constantly released, but understanding what they actually mean for companies or industries requires time and expertise. We wanted to build a system that could convert raw information into structured insights that people can act on quickly. Our goal was to reduce information overload and help users focus on what matters and why it matters.

What it does

Our platform analyzes company-related information and transforms it into concise decision-ready insights. Users first select their interests (for example: AI, finance, technology, energy, or markets). The platform then generates a personalized daily feed with relevant company insights. Each insight includes: A clear summary Key reasoning Possible implications A confidence score Users can also ask questions about specific companies using a built-in chat interface. The system analyzes the request and generates structured insights in under six seconds.

How we built it

The system uses a lightweight architecture designed for speed and efficiency.

Backend

Python

FastAPI

Async processing

AI reasoning

Google Gemini API for analysis

Structured prompting for consistent outputs

To maintain fast performance, we designed a fallback analysis engine. If the AI response takes too long, the system switches to a rule-based heuristic model that analyzes sentiment, keywords, and financial signals to generate insights quickly.

Frontend

React-based interface

Dashboard-style layout inspired by financial news platforms

Chat interface for company questions

Challenges we ran into

The biggest challenge was runtime speed. AI APIs can sometimes take too long to respond, which breaks the user experience. We solved this by implementing a timeout-based fallback system, allowing the platform to automatically switch to a faster analysis method when necessary. Another challenge was balancing accuracy with performance. Heavy machine learning models can improve analysis but dramatically increase runtime and infrastructure cost. Instead, we focused on a hybrid system that combines AI reasoning with lightweight heuristics.

Accomplishments that we're proud of

One accomplishment we’re proud of is building a system that turns large amounts of information into clear, structured insights with reasoning and confidence scores. Instead of just summarizing news, our platform focuses on helping users understand what the information means and what impact it could have. We’re also proud of designing the system to be fast and efficient, returning full analysis in under six seconds while keeping infrastructure lightweight and inexpensive to run. Implementing a fallback analysis engine ensures the platform still produces useful insights even if the AI API is slow. Finally, we’re proud of creating a product that focuses on real-world decision making, combining a personalized news-style dashboard with a chat interface that lets users quickly analyze specific companies or questions.

What we learned

Through this project we learned: How to design AI systems that prioritize speed and reliability How to implement fallback architectures for external APIs How to structure prompts to produce consistent reasoning and confidence scores How to build a product focused on real-world decision making rather than just summarization

What's next for Joblet

Future improvements include: Multi-source analysis from financial news and filings Trend detection across industries Market impact scoring Personalized alerts for major developments Our long-term goal is to create a platform that helps people move from information consumption to intelligent decision-making.