School Project Report - Gruntech Invest

Overview (What the product does)

Gruntech Invest is a web app that helps users compare stocks using a transparent 0–100 fundamentals-based score. It refreshes weekly and highlights top opportunities. For top opportunities, AI is used to produce plain-English explanations and an additional AI score, then a combined score is displayed.

Goal

Help non-experts discover and understand fundamentally strong companies, using clear scoring and simple explanations.

Key idea

Numbers for consistency (score + factor breakdown), AI for readability (short, plain-English summary).

Section A: Using AI to create the project

Criterion A focuses on planning and using AI as a tool. For this project, AI was used as a coding assistant to accelerate development and to help reason about architecture, UI design, and debugging.

A1. First prompt (example)

The first prompt to Cursor would be a short specification of the product. For example:

Create a web app called Gruntech Invest. It should rank stocks with a transparent 0–100 score based on fundamentals. Provide a Top 10 list, stock detail pages, and a clear explanation of why a stock scores well or poorly. Use a weekly refresh pipeline and show data timestamps. Use AI only for writing plain-English summaries for top opportunities.

A2. How AI was used responsibly

AI was used to draft code and propose changes, but every change was verified by running formatting/check scripts and validating outputs.
The scoring rules remain transparent and deterministic; AI does not silently change the score.
When AI services are rate-limited, the app remains usable with the algorithmic score.

Teacher note: The priority is to finish the product first. After the product is stable, document the process (Section B) and reflect on what worked and what did not (Section C).

Section B: Process (What Vincent did to build it)

This section should be written once the product is complete. Below is a clear structure to describe the process.

B1. Planning

Define the audience: investors who want fundamentals, not short-term trading.
Define success: a simple UI that makes the scoring understandable.
Define constraints: weekly refresh, limited API rate limits, predictable costs.

B2. Implementation steps

Frontend: landing page, top lists, stock detail pages, and simple explanations.
Backend: ingestion from data providers, storage in DynamoDB, scoring engine, and API.
Pipeline: scheduled weekly run to refresh data and recompute scores.
Deployment: infrastructure and automated rollouts to the test environment.

B3. Testing and iteration

Validate that top opportunities always include AI scores (when expected).
Fix data issues (for example price chart split artifacts) and confirm no regressions.
Improve clarity: remove confusing text and add tooltips for key metrics.

Section C: Reflection (What worked, what did not)

This section should be written after completion. It focuses on evaluating the outcome and learning from the build.

C1. What worked well

Transparent scoring makes comparisons consistent and explainable.
AI summaries help non-experts understand key reasons and risks faster.
Weekly refresh keeps the platform focused and predictable.

C2. Challenges

Rate limits and AI throttling required resilient fallbacks and retries.
Some tickers have missing fundamentals, so confidence varies by stock.
Keeping UX simple while still educational required careful wording.

C3. Improvements for a future version

More metric explanations directly inside the stock detail view.
Better progress visibility for long-running ingestion and scoring pipelines.
More robust handling for special tickers (preferred shares, ADRs, etc.).

Reflection prompt: “If I had to rebuild this in half the time, what would I simplify? If I had double the time, what would I improve first?”