name: Abdulhaleem Osama location: Dikirnis, Egypt current_role: Software Engineering Lead @ TJM Labs education: B.Sc. Computer & Control Systems Engineering (GPA: 91.8%) Mansoura University (2020-2025) experience: - role: Software Engineering Lead company: TJM Labs period: Sep 2025 - Present highlights: - Leading 20+ automation and AI projects across healthcare clients - Engineering across 3 divisions (Retail, Voice AI, Long-Term Care) - Team of 8 engineers; resource allocation, code review, hiring - Own the Hetzner VPS fleet and internal tooling - role: Software Engineer company: TJM Labs period: Nov 2024 - Sep 2025 highlights: - Joined as 2nd engineer; company since raised $100M, reached 450+ live pharmacies, made 2 acquisitions, automates 500K+ tasks/day - RPA pipelines processing 25,000 prescriptions/day - Voice AI system handling 50% of client calls - ColiVara contributor: eval harness + Python/TypeScript SDKs - role: Data Science Supervisor org: IEEE Mansoura CSC (Student Activity) period: Dec 2023 - Oct 2024 highlights: - Built ML curriculum and mentored committee members - Led technical workshops and training sessions interests: ["MCP & Agent Tooling", "RAG Systems", "NLP", "Healthcare AI", "RPA"]
| Project | Description | Tech | |
|---|---|---|---|
| WhatsApp MCP Server | Self-hosted, authenticated MCP server. Serves 96 tools for the context cost of 29 via progressive disclosure, with a labeled eval proving retrieval routes as well as loading everything. MIT. | Go Python FastMCP Docker |
|
| Alban Osama | Arabic RTL dairy reservation app built and shipped solo for a family business. Live on Google Play. | Flutter Fastify PostgreSQL |
|
| Finance Dashboard | Arabic-first Egyptian market platform: gold, silver, USD parallel, forex, crypto. Append-only partitioned price history; Android app in closed testing. | Next.js TypeScript PostgreSQL |
|
| Healthcare RPA | Automation bots processing 25,000 prescriptions/day across healthcare workflows at TJM Labs. | BotCity Playwright RPA |
|
| AQSA - Arabic NLP Platform | ALLaM Challenge 2024 finalist - Arabic grammar parsing & Quran search with 100% retrieval accuracy. Competed against 177 teams across 17 countries. | LangChain NLP Arabic |
|
| Brain Tumor Classifier | CNN achieving 99.3% accuracy on MRI tumor classification. Live Demo | PyTorch Streamlit CNN |