
Hi! My name is Omar
I build intelligent applications using modern full-stack technologies, machine learning, and scalable backend systems.
Professional Journey
My career timeline and recent professional achievements building production-grade AI and Full-Stack systems.
tp53 health — Building the Cibus Platform
tp53 health builds Cibus, an industry-leading B2B SaaS platform used daily by nutritionists in clinics to manage scheduling, patient records, anamnesis, meal plans, document signing, Gotenberg PDF generation, and automated STS invoicing. Serving over 1,000+ active patient practices, it is a real-world, high-traffic environment where software execution meets medical precision.
Architected Next-Gen AI features including a RAG (Retrieval-Augmented Generation) layer over nutritionists' clinical libraries, streaming AI chats, and Anthropic SDK-style tool integrations with rigid schemas to execute deterministic tasks.
Implemented OpenAI Realtime API integrations for audio and voice-assisted consultation flows. Integrated streaming protocols, advanced prompt caching to slash latency and API costs, and robust server-side utilities.
Shipped core full-stack features using Next.js 15, Prisma, and Postgres (Neon). Engineered advanced patient record workflows, automated PDF generators with Gotenberg, customized Resend transactional mailings, and Stripe billing.
Latest Work
The newest things off my desk — AI products built end to end, from the client through the API to the database.

AI study assistant for Math, Physics, Chemistry and Accounting
A student can type a question, photograph a problem, dictate it aloud, or record a whole lecture, and get a step-by-step explanation back. Answers stream in from a provider-neutral AI layer that routes audio to Gemini and everything else to OpenRouter. It also builds practice quizzes, turns lectures into saved notes, and tracks progress by subject.
Clerk
Express 5
Prisma 7
AI medical scribe for clinicians and small practices
A clinician records a consultation, Scriva transcribes it live with Deepgram, then drafts a structured clinical note from the transcript with Claude and files the note, transcript and generated PDFs under the right patient. Around that sits a practice workspace: patients, intake sheets, Google Calendar sync, branded PDF reports, and an assistant that answers questions grounded in the clinician's own notes.
Prisma 6
NextAuth 5Portfolio
A selection of my recent work showcasing full-stack development, AI integration, and scalable architectures.

A modern SaaS landing page engineered with advanced animations and micro-interactions to deliver a premium user experience. Built with a scalable full-stack architecture, it showcases high-performance frontend design, smooth motion systems, and conversion-focused UI patterns used in modern SaaS platforms.
Framer Motion
Machine learning system to detect fraudulent Medicare healthcare providers. Performs feature engineering across 4 data sources, handles class imbalance, and evaluates multiple models — achieving 93% accuracy and 0.93 ROC-AUC with Gradient Boosting.

AI-driven meal planning platform. Generates personalized meal plans based on user preferences and dietary needs. Showcases intelligent algorithm design, React frontend, and optimized state management for smooth UX.
Clerk
AI-powered code tutorial generator. Upload a code snippet and get a detailed, step-by-step AI-generated tutorial. Built with a focus on clean architecture, efficient data handling, and seamless developer experience.
PostgreSQL
Prisma
ClerkSkills & Tools
Technologies and tools I use to build scalable, modern applications
React
Next.js
TypeScript
Python
JavaScript
Tailwind
CSS
React Query
Notion
Material-UI
Node.js
Express
Java
PostgreSQL
MongoDB
Redis

Prisma
Supabase
HTML
Git
GitHub
React Native
Vercel
Docker
AWS
Cloudflare
Linux
Postman
Apache Kafka
Socket.io
Framer Motion
Redux
Jest
GitHub Actions
Sentry
Zustand
tRPC