Production AI systems · Multi-agent platforms · Computer vision · Cloud infrastructure · Reliable delivery
I am an AI & DevOps Engineer based in Cairo, Egypt. I work across the full delivery path—from model evaluation, RAG pipelines, and multi-agent orchestration to backend platforms, containers, observability, CI/CD, and secure cloud infrastructure.
My focus is not only making an AI feature work. I build the engineering around it so it can be tested, deployed, monitored, recovered, and improved safely.
أنا عبدالهادي علي، مهندس ذكاء اصطناعي وDevOps. أبني أنظمة ذكاء اصطناعي جاهزة للإنتاج، ومنصات متعددة الوكلاء، وحلول رؤية حاسوبية، وبنية تحتية موثوقة من النموذج وحتى النشر والمراقبة.
- AI Engineer at AIVIS Studio, building production computer-vision systems for edge devices
- Former AI & DevOps Engineer at Zedny INC, working on internal platforms, multi-agent systems, and infrastructure reliability
- Bachelor's degree in Computer Science and Artificial Intelligence from Benha University
- Open to full-time roles and selected freelance work in AI systems, computer vision, backend engineering, MLOps, and DevOps
| Project | What I engineered | Selected result |
|---|---|---|
| Real-Time PPE Compliance Detection | Edge computer vision across ten object and compliance-state classes using YOLO11, OWLv2, supervised CLIP, TensorRT, and ByteTrack | Cut false positives by 95% and raised annotation F1 from 0.39 to 0.81 |
| Autonomous DevOps Agent SDK | Provider-neutral multi-agent SDK for deployment, monitoring, RAG-assisted incident investigation, and controlled remediation | One architecture spanning CI/CD, observability, knowledge, remediation, and alerting agents |
| Internal DevOps Dashboard | Secure operations platform for VMs, Kubernetes, monitoring, audit history, and DigitalOcean drift detection | Unified infrastructure and observability in one operational interface |
| Interactive Course Agents | LangGraph platform generating matched visuals, layouts, overlays, and self-correcting Manim animations | Preserved 1,300+ passing tests while driving broken animation renders to approximately zero |
| CodeFlow | AI code-intelligence platform for source-cited RAG, living documentation, architecture diagrams, and agent-assisted review | Turned repositories into navigable, versioned knowledge systems |
| MAISYS | Bilingual medical AI platform with 13 microservices, hybrid retrieval, safety controls, and AWS infrastructure | Reduced measured hallucination rate from 14.7% to 3.2% |
| Infrastructure & Reliability | Production recovery, isolated environments, secure routing, observability, and self-healing deployment pipelines | Built verification, migration, health-gate, and automatic rollback workflows |
| PaperFlow | Microservices-based research assistant for grounded paper chat, discovery, summarization, translation, OCR, and voice | Eight specialized services behind a unified API gateway |
Selected implementations are private. The portfolio presents approved architecture, engineering decisions, demonstrations, and outcomes without exposing confidential source code.
AI, agents, and vision
RAG · LangGraph · CrewAI · LiteLLM · YOLO11 · OpenCLIP · Model evaluation · Multi-agent systems
Backend and data
Async APIs · WebSockets · SQLAlchemy · Alembic · Qdrant · pgvector · MongoDB · MinIO / S3
Cloud, platform, and reliability
Linux · nginx · Loki · Alertmanager · Cloudflare · CI/CD · Automated rollback · Secrets management
If you are working on production AI, computer vision, agent platforms, backend systems, or cloud infrastructure, I would be glad to connect.
- Portfolio: abdelhady-portfolio-pearl.vercel.app
- LinkedIn: abdelhady-ali-940761316
- Email: abdulhadi.ali.mhmd@gmail.com
Build clearly. Deploy safely. Improve continuously.