Building reliable LLM systems, autonomous coding agents, and agentic infrastructure.
Agentic AI Intern at Sofrecom Tunisia · AI Engineering at ESPRIT · Expected 2027
The projects that best represent how I currently think about AI engineering: autonomy, reliability, structured execution, and developer tooling.
Stack: Python · Asyncio · LLM Agents · LSP · DAP · DevTools
A provider-agnostic terminal coding agent built for long-horizon repository work. Grinta plans tasks, edits code, executes commands, observes failures, validates outcomes, and recovers across interrupted sessions.
- Event-stream ledger, checkpoints, replay/recovery, and state reconstruction
- Structured Chat / Plan / Agent execution modes with configurable autonomy
- Local LSP + DAP integration for code navigation and interactive debugging
- Context compaction, stuck-loop detection, recovery flows, and completion validation
- OpenAI-compatible APIs, Anthropic, Gemini, OpenRouter, Ollama, and LM Studio
- Public release candidate with cross-platform CI and documented security boundaries
🔗 Repository · 🎥 Demo · 📖 Architecture
Stack: Next.js · TypeScript · Zod · Gemini · SSE · Prisma
A structured multi-agent platform coordinating seven AI roles across product planning, architecture, infrastructure, security, UI/UX, implementation, and QA.
- Schema-driven agent outputs rendered as interactive React artifacts
- Targeted artifact refinement with operations such as
set_at_pathandadd_array_item - SSE-based job progress for long-running generation workflows
- Gemini context caching to reduce repeated prompt overhead
- Export pipelines for Markdown, PDF, HTML, and PPTX
Stack: Python · LlamaIndex · Pinecone · RAG · Regulatory AI
A compliance parsing workflow for financial documents that extracts ESG-related data patterns and checks them against Sustainable Finance Disclosure Regulation requirements, paired with enterprise retrieval workflows for unstructured knowledge-base queries.
🔗 SFDR Validator · Enterprise Knowledge Base
Stack: Python · LLM Architecture · Tool Use · MCP · Evaluation
A clean-room educational reference library covering 21 agent workflow primitives: routing, planning, reflection, tool use, multi-agent coordination, memory, MCP, guardrails, evaluation, and more.
Currently working on enterprise agentic AI around autonomous data-reliability workflows.
My work focuses on deterministic extraction, semantic data alignment, schema-drift detection, controlled ETL repair, sandboxed Python/SQL execution, confidence-based action thresholds, and validation-oriented agent control flows.
|
Distributed orchestration with LangGraph state workflows, Redis coordination, and real-time communication. LangGraph · FastAPI · Redis · WebSockets → Platform · Orchestrator |
Observability and evaluation utilities for agent workflows: inference cost, latency, token density, and LLM-as-Judge fixtures. LangSmith · Pytest · LLM Evaluation → Observability · Evaluation |
|
AI-powered UI generation workflow that turns natural-language requests into structured React interfaces. Next.js · Gemini · shadcn/ui → Repo |
Integration layer for connecting agent workflows with external communication and service surfaces. Agents · Integrations · APIs → Repo |
|
Multi-module AI ecosystem spanning art authentication, generation, restoration, 2D→3D, and RAG Q&A. PyTorch · Stable Diffusion XL · Vision Transformers → Repo |
Platform for analyzing image, audio, and video inputs with multimodal AI workflows. Vision · Audio · FastAPI → Repo |
|
Agent-based support workflow with ticket routing, escalation, and QA roles. CrewAI · FastAPI · SQLAlchemy → Repo |
Collaborative research workflow combining retrieval, web research, fact-checking, and citation generation. LangChain · FAISS · Streamlit → Repo |
|
Autonomous goal decomposition and adaptive task execution reference project. LangChain · Agents · Streamlit → Repo |
Full-stack healthcare management platform with RBAC and analytics workflows. Django · Power BI · SQLite → Repo |
AI Agents & LLM Systems: Tool Calling · MCP · ReAct · Planning · Context Management · LangGraph · LlamaIndex · RAG · LLM Evaluation · Agent Observability
Backend & Infrastructure: Python · Asyncio · FastAPI · Subprocess Runtimes · Docker · Redis · SQLite · PostgreSQL · REST APIs · SSE · WebSockets
Frontend & Software Engineering: TypeScript · React · Next.js · Java · Spring Boot · SQL · Git · Tailwind CSS
AI & Data: Vector Search · Pinecone · ChromaDB · FAISS · OCR · Anomaly Detection · Multimodal AI
Languages: Arabic (Native) · English (Fluent) · French (Fluent)
- 🤖 Autonomous agent runtimes with tools, recovery, state, and validation
- 🧰 Developer tooling for code navigation, execution, and debugging
- 📊 Agent reliability systems for evaluation, observability, and completion integrity
- 🔍 Enterprise AI workflows spanning RAG, data alignment, OCR, and controlled automation
- 🏗️ Full-stack AI products from backend orchestration to interactive user interfaces
Open to a 26-week final-year engineering internship starting in early 2027.
AI Agents · LLM Infrastructure · Developer Tools · Applied Software Engineering