Investor Evolution

Personalized Investment Portfolio News Tracker. For everyday busy investors.

Comment

Inspiration

We are developers who like investing, but without being professional investors, it's hard to devote the time it takes to do the analysis. We want to make our lives easier and leverage the power of AI.

What it does

You can get recent news summary specific to your portfolio sent straight to your inbox

How we built it

We use third-party API to fetch news articles regarding the stocks, then convert them into vectors and store them in Pinecone with namespace as "news" because we know that stock analysis goes beyond qualitative data, so if in the future, we want to build on top of this. We can have specific queries for specific features. Summarization is done with langchain and openai. All of those actions lives on a EC2 instance as an API. We have another piece with Nextjs that handles the frontend and interacting with supabase data. Data from supabase are directly user-related and its realtime database helped us give user feedback on the progress of the summarization

Challenges we ran into

There are a lot of challenges. The primary challenge was fetching articles, storing vectors, and summarizing can take a bit of time. That's why we have those functions on EC2. Initially, we tried Vercel's edge function, serverless Lambda function. Took a big chunk of our time rewriting and adapting. Also, initially, we wanted to use Plaid to let user import their portfolio, but we found out that it's quite a bit of a process to gain production access.

Accomplishments that we're proud of

Overcoming a lot of last-minute changes. Working with ChatGPT to learn faster and build faster.

What we learned

The plan never goes according to plan. Need to have back-up plan on back-up plan

What's next for Investor Evolution

We want to get users and create a feedback loop where like-minded investors will help us build tools for them.

Built With

Share this project:

Updates