Abdellatif A.

Machine Learning | Deep Learning Engineer | AI Services

I build production-grade AI systems, not prototypes that stall after the demo. My focus is LangChain, LangSmith, RAG pipelines, and MCP-based agentic workflows, taken from architecture through deployment and monitoring. If your project involves multiple moving parts, data pipelines, tool-calling agents, retrieval systems, APIs, that need to work together reliably at scale, that's where I do my best work. Core expertise: LLM & Agentic AI — LangChain, LangSmith, LangGraph, MCP, tool and function calling, multi-agent workflows. LLM providers — Claude API and Anthropic, OpenAI, and other major LLM providers, with experience routing tasks across models based on cost, latency, and reasoning requirements. RAG & Retrieval — embeddings, chunking strategies, vector databases including Pinecone, Qdrant, Weaviate, and Milvus. Automation & orchestration — n8n and workflow automation platforms, connecting AI agents to real business systems and APIs. Production infrastructure — FastAPI, async Python, Docker, CI/CD, tracing and evaluation with LangSmith. Cloud & MLOps — AWS, Azure, and GCP deployment, Kubernetes, model monitoring, cost-optimized inference. Deep learning foundations — PyTorch, TensorFlow, transformer fine-tuning, computer vision with YOLO and Vision Transformers when a project calls for it. I've architected and delivered end-to-end AI systems across LLM applications, computer vision, and NLP, with a consistent focus on debugging complex production issues, memory handling, async concurrency, malformed tool outputs, that only surface under real load, not in notebooks. I communicate clearly, document what I build, and deliver on time. If you have a complex, production-bound AI project, let's talk about it.