Houssem Eddine
M.Eng. AI @Hochschule Schmalkalden | Werkstudent / HiWi in AI Engineering | LLM · RAG · Agentic Systems | PyTorch · LangChain · Azure · MONAI
About
AI Engineer building production-ready LLM, RAG, and Agentic AI systems that reach deployment, not just notebooks. Pursuing M.Eng. in Artificial Intelligence at Hochschule Schmalkalden (Erasmus+, WS 2026/27), working full-time as AI Engineer at Alight Consulting GmbH in Germany. Open to Werkstudent and HiWi roles in Thuringia and remote. What I have shipped: Delivered a production-grade Document AI RAG platform on Microsoft Azure: hybrid BM25 + semantic retrieval (Azure AI Search), structured document extraction (Azure Document Intelligence), HyDE query expansion, LLMLingua context compression, server-side RBAC with multi-tenant document isolation, and a Next.js + FastAPI application with source-cited answers and persistent conversation management. Built a clinical-grade 3D cardiac segmentation pipeline (MONAI, PyTorch, Dice 0.88-0.91, rated clinically acceptable in 95%+ of blinded cardiologist evaluations, EU MDR and IVDR compliant) and fine-tuned a clinical LLM with SFT and DPO, cutting hallucination rate by 35% and reporting time by 30%. Built a multi-agent defect detection system at SBT Tunisia using vision and language models with a custom MCP server, reducing automotive component return rates from 6.8% to 0.2% on a live assembly line. Built an agentic GTM intelligence pipeline targeting the French and Italian automotive supply chain: automated signal detection across job postings, supplier sites, and conferences to identify Tier-1 wire harness buyers before procurement cycles open, delivered as a live reporting dashboard. Published at KES 2026: multi-agent VLM pipeline for brain tumor analysis. 1st place Mutualhack 3.0 (EduVerse, multimodal multi-agent adaptive learning). 2nd place ESPRIT x Recordati (GreenOPS AI, pharma energy optimization, human-in-the-loop). Domains: Medical AI, Computer Vision, NLP, Deep Learning, MLOps, Agentic AI. Regulated environments: ISO 13485, EU MDR, IVDR, industrial manufacturing. Stack: Python, PyTorch, MONAI, LangChain, LangGraph, Hugging Face Transformers, Microsoft Azure (AI Search, Document Intelligence, OpenAI, Blob Storage), YOLOv8, ONNX, Docker, FastAPI, Neo4j, MCP (Model Context Protocol), Git. Fine-tuning: LoRA, QLoRA, SFT, DPO, RLHF. Multi-agent orchestration: CrewAI, LangGraph. Languages: English (fluent), French (professional), German (B1, targeting B2). Schmalkalden, Germany.
Company history
- NewrestData Analyst2023-06 – 2023-07 · 1 mo
- NewrestData Analyst – Forecasting & Supply Chain Automation2023-07 – 2023-09 · 2 mo
- SBTArtificial Intelligence Engineer2026-01 – 2026-06 · 5 mo
- SBTAI GTM Engineer2026-06 – present · 2 mo
- Alight Consulting GmbHArtificial Intelligence Engineer2026-06 – present · 2 mo
- Alight Consulting GmbHMedical AI Engineer2025-06 – 2025-09 · 3 mo
- RecordatiAI Engineer2026-05 – 2026-06 · 1 mo
- DeepFlowTechnical Manager2024-09 – 2025-01 · 4 mo
- DeepFlowEducation Mentor2023-06 – 2025-06 · 2y
- CMA CGMArtificial Intelligence Engineer2025-01 – 2025-06 · 5 mo
- Self-employedFounder2023-05 – present · 3y 3mo
Skills
Playbooks
All playbooks →No playbooks indexed for this profile yet.
Redo