I buildRAG Pipelinesfor production.
I help businesses eliminate manual processes, cut costs, and scale faster — with custom AI built for production, not just demos.
- Save 10–30h/week automating document & data tasks
- Cut AI costs by up to 73% with production-grade caching
- Go live in weeks, not months — fully tested
User Query
Unstructured Docs
Embedding & Chunking
sentence-transformers · 512 tok
Qdrant + Hybrid Search
BM25 · Dense vectors · Top-5
Claude / LLM Synthesizer
Production Output · Azure
Automation Specialist · Morocco
Ayoub Habib Eddine
Automation Specialist
My background isn't CS - it's logistics. Six years co-running a moving company taught me that a workflow is only good if it survives real failure. Now I build production-grade AI: RAG pipelines, LLM platforms, and automation at DXC Technology and Ark-x.
Why work with me
Problems I solve.
Incidents written on paper, hard to find at audit time
Every case recorded in under two minutes, with a complete history
Full history in under 2 minKPIs rebuilt by hand in Excel each month
Indicators calculated automatically from the records
Auto-calculated, always liveStaff interrupt a manager to ask which procedure applies
An assistant answers from your own documents and cites the source
AI answers from your own docsSensitive data that cannot go to the cloud
A system hosted on your premises, with only anonymised text sent to AI services
Local model, zero cloud exposureDeadlines for maintenance and training tracked from memory
Automatic reminders before each due date
Alerts sent before each due dateMy services.
From RAG pipelines to n8n workflows - systems engineered for reliability, not just demos.
RAG Systems
Production-grade retrieval-augmented generation pipelines grounded in your company's real data.
- Document ingestion and chunking pipeline
- Vector search with source citations
- Evaluation and quality monitoring
Workflow Automation
End-to-end automation with n8n and Make — cutting repetitive manual work for your team.
- Process mapping before any tool is chosen
- Trigger-based pipelines with error handling
- Team training and documentation
API Orchestration
Multi-service integrations: FastAPI backends, webhook handlers, and reliable data pipelines.
- FastAPI service design and deployment
- Webhook and event-driven architecture
- Logging, retries, and failure recovery
AI Agent Development
LangGraph-powered agents with memory, planning, and tool use — deployed on Azure with Docker.
- Agent architecture and tool definitions
- Memory and state management
- Containerised deployment with health checks
Knowledge Graphs
Graph-RAG with Neo4j that surfaces relationships across documents and enterprise data sources.
- Entity and relationship extraction
- Graph schema design
- Hybrid vector + graph search
Voice AI Agents
Conversational voice agents that handle real calls — powered by ElevenLabs, Vapi, and LangGraph with human-like latency.
- End-to-end voice pipeline (STT → LLM → TTS)
- Telephony integration via Vapi or Twilio
- Custom voice cloning and persona design
Recent work.
Production systems at DXC Technology, Ark-x, and beyond.
LLM cost reduction
AI Watch - Enterprise RAG Platform
Full-stack LLM platform with automated document ingestion, Azure OpenAI integration, and knowledge graph synthesis. Two-tier caching (Redis + Supabase) cut API costs by 73%.
ComGenAI - Communication Generator
RAG-based system generating enterprise communications from unstructured documents. 8-service Docker stack on Azure VM with auto-TLS and health checks.
Ahlan - Multimodal Hospitality Concierge
Voice + text AI concierge for hospitality, used as a flagship enterprise demo. Handles check-ins, recommendations, and guest requests autonomously.
Moroccan Constitution Chatbot
RAG pipeline grounding an LLM in the full Moroccan Constitution. Answers legal questions with cited articles, deployed as a conversational interface.
Intelligent ATS Automation System
Hackathon project automating resume screening and candidate scoring. Led the team to build an end-to-end pipeline integrating job descriptions with candidate profiles.
LinkedIn Content Generator
Automated content pipeline generating AI-themed LinkedIn posts, scheduling them, and tracking engagement - zero manual effort after setup.
AI-Assisted Quality Management System
Local web application covering non-conformity register, KPI dashboard, purchasing, staff onboarding, regulatory watch and Excel export. AI layer runs behind a privacy gateway.
My method
How I work.
Start on site
I visit the business, follow real cases from start to finish, and note where time is lost.
Map the work before the tools
Each repetitive task is written down: what triggers it, who does it, how often, and what goes wrong.
Show a working demo within days
You click through your own screens with sample data before committing to anything.
Build in phases
One useful module at a time, each validated by your team before the next begins.
Stay after delivery
Updates, backups and support are handled remotely, with on-site visits only when needed.
Start on site
I visit the business, follow real cases from start to finish, and note where time is lost.
Map the work before the tools
Each repetitive task is written down: what triggers it, who does it, how often, and what goes wrong.
Show a working demo within days
You click through your own screens with sample data before committing to anything.
Build in phases
One useful module at a time, each validated by your team before the next begins.
Stay after delivery
Updates, backups and support are handled remotely, with on-site visits only when needed.
Automation Specialist · Morocco
Ayoub Habib Eddine.
My background isn't CS - it's logistics. Six years co-running a moving company taught me that a workflow is only good if it survives real failure. Now I build production-grade AI: RAG pipelines, LLM platforms, and automation at DXC Technology and Ark-x.
AI & LLM
Data Storage
Infrastructure
What they say.
From colleagues and clients.
“J'ai travaillé avec Ayoub sur plusieurs projets et c'est quelqu'un sur qui on peut compter. Sérieux, efficace, et surtout curieux : il apprend vite et se tient toujours au courant des dernières nouveautés en IA. En peu de temps, j'ai vu ses connaissances évoluer de façon impressionnante. Je le recommande les yeux fermés — il apportera beaucoup à n'importe quelle équipe.”
“I highly recommend Ssi Ayoub. The team was very professional, punctual, and organized from start to finish.”
“Thank you so much Ayoub and your team! You're serious, punctual people who do efficient and professional work.”
Frequently asked questions.
Quick estimate
What is it worth to you?
Assumption: 70% automation rate · €25/hr
hours saved / month
estimated value / month