RAG Knowledge Assistant
Production-ready retrieval pipeline with hybrid search, reranking, and cited responses. Handles multi-format ingestion and streams answers under 200ms TTFT.
PythonLangChainQdrantFastAPIOpenAI
available for new projects
Kareem Almasri
< />
I build production-grade LLM applications — retrieval pipelines, agentic workflows, and fine-tuned models that ship real value. Currently exploring the edges of context engineering and multi-modal RAG.
Kareem@portfolio: ~
❯
01. stack
Tools that ship.
A pragmatic toolkit for taking LLM ideas from notebook to production — model orchestration, retrieval, evals, and the infrastructure to glue it all together.
# core
PythonFastAPIDocker
# llm
LangChainRAG SystemsVector DatabasesPrompt EngineeringFine-tuning (LoRA/QLoRA)
# providers
OpenAI APIAnthropic APIHuggingFace
02. work
Selected projects.
A few things I've built. Each one solved a real problem and taught me something I now bring to the next one.
Agentic Workflow Engine
Multi-step agent runtime with tool calling, memory, and human-in-the-loop checkpoints. Built for reliability with full trace logging and replay.
PythonAnthropicPydanticRedisDocker
03. contact
Let's build something.
Working on an LLM problem? Want to collaborate? Drop a message and I'll get back to you within a day.