The “Recruiter’s Eye”: The app doesn’t just scan for words; it thinks like a human recruiter. It spots where your experience is strong and where you might be leaving out the “proof” that hiring managers look for.
ATS Scorecard: Different companies use different software to screen resumes. Our tool shows you how you rank across the three most popular systems (Greenhouse, Lever, and Workday) so you know exactly where you stand.
Evidence-Based Matching: Instead of guessing, the AI highlights the exact sentences in your resume that prove you can do the job. If it can’t find proof, it tells you exactly what’s missing.
Smart Rewriting: It provides a “Before and After” for your most important bullet points, showing you how to rephrase your achievements to sound more impactful without changing the facts.
One-Click Analysis: Simply paste a link to your Google Doc resume and the job you want. In seconds, you get a full strategic report on how to win the interview.
Professional Portfolio Analysis
Key Features of the Approach:
By using Google AI Studios I was able to create an APP less than 20 minutes. The code was generated using AI and then fine tuned with AI input. 🚀
01.
Project Synopsis: AI Powered Portfolio Analysis
This project is an AI-powered Investment Statement Analysis Agent designed to transform complex, unstructured brokerage statements into professional, standardized portfolio reports.
Core AI Logic & Architecture
The application leverages the Gemini 3.1 Pro model to perform sophisticated financial data extraction and analysis. The AI logic is structured around several key pillars:
Multimodal Data Processing: The system accepts both raw images and extracted text from brokerage statements. The AI uses these inputs in tandem, utilizing the images for visual context and OCR fallback to ensure high-fidelity data capture from complex table layouts.
Strict Schema Enforcement: To ensure the UI can reliably render charts and tables, the AI logic utilizes Structured Output (JSON Schema). This forces the model to map disparate statement formats into a unified data model containing:
Metadata: Account identification, period detection, and confidence scoring.
Financial Quantities: Precise extraction of beginning/ending values, net changes, fees, and dividends.
Portfolio Composition: Detailed holdings analysis, including ticker symbols, market values, and asset class categorization.
Intelligent Classification & Inference:
Statement Type Detection: The logic automatically distinguishes between “Monthly” and “Year-End” reports, adjusting the depth of analysis accordingly.
Style & Asset Class Mapping: The AI identifies common ETFs and mutual funds to infer investment styles (e.g., Large Cap Growth) and asset classes (e.g., Fixed Income) even when not explicitly stated.
Predictive & Risk Modeling: For year-end reviews, the AI logic generates 10-year growth projections across multiple return variables and simulates risk scenarios (e.g., Bear Market, Severe Recession) based on the detected portfolio beta and composition.
Contextual Commentary: Beyond data extraction, the model generates professional wealth-management commentary, synthesizing “Executive Summaries” and “Performance Drivers” that adapt to a user-selected tone (e.g., Conservative, Balanced, or Growth-oriented).
Safety & Privacy Guardrails: The system instructions mandate the masking of sensitive account information and enforce strict markdown sanitization to prevent “hallucinated” styling or HTML injections, ensuring a secure and consistent user experience.
This ATS Scorecard is provided for educational and demonstration purposes only. Results are automated estimates and do not guarantee hiring outcomes or reflect any specific employer’s actual screening process. The creator assumes no responsibility for decisions made based on this tool’s output.
This project was developed as a portfolio demonstration showcasing coursework completed through Johns Hopkins University in Baltimore, Maryland.