🧑‍💻 AI Engineer · 🤖 ML Systems · 🏗️ Agent Architectures · ⚙️ Product Management

Building production AI systems and multi-agent workflows. Backend-first, clean architecture, scalable by design.

CONNECT WITH ME:

Hi! I'm Anas. I'm a backend developer who blends systems thinking with AI/ML. I specialize in building reliable, scalable systems with a strong focus on multi-agent workflows, LLM automation, and clean architecture. While I know my way around frontend, what truly drives me is designing backend infrastructure and production-ready AI pipelines.

Currently based in Houston, TX and building the AI agent platform at Katy ISD (80+ campuses), while scaling MedGrid, an AI-powered medical education platform serving 1,200+ students.

• 🚀 Shaping the future of tech: building systems and tools that push what's possible
• 🤖 AI systems & agent frameworks: composable agents, clear failure boundaries
• 📐 System design: thinking in components, data flows, and failure modes
• 🏗️ Clean backend architecture: modular, testable, maintainable
• 🛠️ Building internal tools and automation that solve real problems
• 💻 Workflow optimization: if it can be automated, it will be
• 🔧 Long-term maintainability: quality over quick wins
• 🌱 Growing in public: building toward meaningful open source contributions

💡 I'm proficient with frontend technologies and modern build tools, but my true passion lies in backend and AI systems: where structure, logic, and performance matter most. While I can create polished user interfaces and work with frameworks like React and Next.js, what really excites me is architecting robust server-side infrastructure, designing multi-agent workflows, and building production ML pipelines that scale.

Note: I also work with AI-assisted development tools (Cursor, Claude, GitHub Copilot) and enjoy hands-on experimentation with embeddings, RAG systems, and LLM automation.

Project	Description	Status
HekimaRisk	AI-native geopolitical & supply chain risk intelligence platform for Africa-focused decision-makers	🚧 In Progress
MedGrid	AI-powered medical education platform with RAG-driven adaptive learning, 1,200+ students medgriddy.com	🚧 In Progress
Katy ISD	District-wide AI agent platform: multi-agent call routing, building kiosks, drone surveillance	🚧 In Progress

• React (Meta) – Fixed regression in concurrent rendering for edge cases
• LangChain – Added support for streaming in SQL agent
• TensorFlow (Google) – Fixed gradient computation for custom layers in TF 2.x
• RAPIDS (Nvidia) – Optimized GPU-accelerated DataFrame operations in cuDF
• Transformers (Hugging Face) – Added tokenizer support for new model architectures
• MLflow (Databricks) – Improved experiment tracking API for distributed runs
• Chaos Monkey (Netflix) – Enhanced resilience testing for microservice failure scenarios
• FastAPI – Improved OpenAPI schema generation for nested Pydantic models
• TypeScript (Microsoft) – Fixed type inference for conditional mapped types
• Kubernetes (CNCF) – Documentation and Helm chart fixes for ingress controller
• Pandas (NumFOCUS) – Performance fix for groupby on nullable dtypes

• Perseus Health AI – AI Voice Agents & Chatbots for healthcare. Streamlining patient interactions, appointment scheduling, and workflows for medical facilities.
• 🤖 Enterprise Intelligence Agent – Backend AI chatbot connecting enterprise data, SQL analytics, and ML churn prediction. FastAPI, OpenAI tool-calling, scikit-learn, Docker.
• 📍 PinDrop – AI study assistant Chrome extension. GPT-4o-powered hints, explanations, and reasoning for highlighted study content. Node.js backend + extension frontend.
• 📋 Feature Prioritizer – Interactive web app for data-driven product feature prioritization using RICE and MoSCoW frameworks. Next.js, TypeScript, Tailwind CSS, Chart.js.

💬 Thanks for stopping by. Always open to interesting conversations about backend, AI systems, or building things the right way.