indigo-resume

byRamji Vaghani

1. System OverviewA SaaS platform that leverages AI to instantly tailor a user’s base resume. It operates via two primary data ingestion modes to output highly optimized, ATS-friendly resumes (PDF/DOCX) targeted at specific roles or industry standards.2. Core WorkflowsMode A: LinkedIn URL + Base ResumeUser uploads a base resume (PDF/DOCX) and pastes their LinkedIn profile URL.Backend triggers the Bright Data API to scrape the public LinkedIn profile for recent endorsements, skills, and updated job descriptions.The AI engine cross-references the base resume with the scraped LinkedIn data to fill gaps, update timelines, and modernize the phrasing.Mode B: Job Description (JD) + Base ResumeUser uploads a base resume and pastes the text/URL of a target Job Description.The AI engine extracts hard skills, soft skills, and core requirements from the JD.The AI rewrites the base resume’s bullet points to mirror the JD’s language, optimizing for Applicant Tracking Systems (ATS) without fabricating experience.3. Technical ArchitectureFrontend (User Interface): React.js or Next.js (for fast rendering and easy state management for document previews).Backend (API & Logic): Python (FastAPI) or Node.js (Express). Python is highly recommended for seamless AI and data processing integration.Database: PostgreSQL (relational data for users, transaction history) and AWS S3 / Google Cloud Storage (for securely storing encrypted resume files).External APIs:Bright Data: For bypassing anti-bot measures and scraping LinkedIn profiles.LLM Engine: OpenAI (GPT-4o) or Google Gemini for natural language processing and document restructuring.Document Generation: CloudConvert or standard Python libraries (python-docx, pdfkit) to render the final tailored resume.4. User Authentication & SecurityLogin Methods: Google OAuth, LinkedIn OAuth (reduces friction), and standard Email/Password.Session Management: JWT (JSON Web Tokens) for secure, stateless API communication.Data Privacy: Uploaded resumes must be automatically deleted from servers after 30 days. PII (Personally Identifiable Information) must be encrypted at rest.5. Pricing & Monetization (Stripe Integration)A hybrid credit-based and subscription model works best for this user behavior.TierPriceFeaturesTarget AudienceBasic (Free)$01 Tailored Resume via JD mode, Watermarked output.First-time users testing the AI quality.Pay-As-You-Go$3 / credit1 Credit = 1 Tailored Resume (JD or LinkedIn mode). No watermarks.Passive job seekers applying occasionally.Pro (Monthly)$15 / monthUnlimited JD tailoring, 20 LinkedIn scrapes/mo, Cover Letter generation.Active job seekers applying in high volume.6. The AI Data Pipeline (Step-by-Step)Parsing: Extract raw text from the uploaded PDF/DOCX using OCR or text-extraction libraries.Structuring: The LLM converts the unstructured resume text into a structured JSON format (Experience, Education, Skills).Analysis (The "Tailoring"): * If JD: LLM scores the JSON resume against the JD. It identifies missing keywords and rewrites experience bullets to highlight overlapping skills.If LinkedIn: LLM merges the JSON resume with Bright Data's JSON output, prioritizing the most recent or detailed descriptions.Formatting: The updated JSON is injected back into a standardized, ATS-compliant document template.Delivery: The system compiles the final file and provides a download link to the user interface.