Full Stack Developer
The Right Contact
AI-powered contact lens knowledge hub — natural language search, structured data lookups, and corneal image analysis for eye care professionals
Overview
TheRightContact.com is an advanced knowledge platform built to help optometrists and eye care professionals access contact lens information quickly and accurately. Unlike generic medical databases, it is highly specialized — combining AI-driven natural language search, structured Excel-style lookups, and image recognition for corneal topography analysis. The platform was created to eliminate the inefficiency of combing through dense vendor PDFs and manuals. Eye care professionals can query lens specifications, patient compatibility criteria, and prescription details in plain language, getting clinically reliable answers in seconds from a continuously updated knowledge base of 513+ trained documents.
The Challenge
Domain-specific accuracy was the core challenge — the AI needed to provide clinically reliable answers, not generic responses. This required training on optometry-specific datasets and tuning GPT for terminology like toric lenses, base curves, and prescription ranges. Corneal topography scans vary greatly in quality and format, requiring preprocessing pipelines to normalize inputs before image analysis. The contact lens industry also evolves rapidly, so the platform needed an admin-friendly update pipeline to keep specifications current without developer intervention.
How We Solved It
Structured Data Modeling
Developed a database architecture to handle vendor specifications, prescription ranges, and lens compatibility guidelines in a structured, Excel-like lookup format. Eye care professionals can filter and search across 513+ documents with precision — not just keyword matching.
AI-Powered Natural Language Search
Integrated OpenAI GPT-4 tuned for optometry terminology to understand domain-specific queries ('What toric lenses are available for this prescription range?'). The AI draws from the structured knowledge base to return accurate, contextually appropriate answers.
Computer Vision — Corneal Topography Analysis
Added image recognition capabilities to interpret corneal topography scans. After preprocessing to normalize quality and format variations, the system provides AI-assisted lens compatibility suggestions directly from the scan data — removing manual interpretation steps from clinical workflows.
Excel-Style Data Lookup Interface
Built a structured query interface mirroring familiar spreadsheet-style filtering. Professionals can search by lens type, manufacturer, prescription parameter, or compatibility criteria — combining AI search with deterministic structured lookups for maximum accuracy.
Continuous Knowledge Base Updates
Created an admin system for uploading new lens data, PDFs, and Excel sheets directly into the training pipeline. The Training Log dashboard tracks all 513+ ingested documents with author, publish date, and training timestamp — ensuring the knowledge base grows alongside the industry.
Scalable Cloud Deployment
Deployed on AWS with Redis caching for fast repeated queries. The platform is fully responsive — accessible from clinic desktops and mobile devices. PostgreSQL stores structured records while Redis handles session caching for low-latency search responses.
Platform
Results & Impact
513+ Trained Documents
Comprehensive knowledge base of vendor PDFs, Excel sheets, and clinical guidelines — all ingested, indexed, and searchable via natural language.
Instant Clinical Answers
Optometrists get accurate, domain-specific lens recommendations in seconds — eliminating time spent searching through vendor manuals.
AI + Structured Search
Dual-mode access: natural language queries via GPT-4 and precision Excel-style filtering — giving professionals both flexibility and accuracy.
Corneal Topography Analysis
Computer vision interprets topography scans and suggests compatible lens options, adding AI-assisted clinical decision support to the workflow.
Admin-Managed Updates
Non-technical admins can upload new lens data and documents through the dashboard — keeping the platform current without engineering work.
Multi-Device Accessible
Fully responsive platform accessible from clinic desktop workstations and mobile devices — usable at the point of care.
Technical Architecture
Have a similar project?
I'm open to freelance projects and full-time roles. Let's build something great together.