Hamed Khademi Khaledi
AI Engineer & NLP Researcher
Professional Summary
AI Engineer and NLP Researcher experienced in building and deploying Large Language Models (LLMs), LLM-based agents, Retrieval-Augmented Generation (RAG) systems, and modular NLP microservices. Skilled in Python, PyTorch, vLLM, LangChain, and LangGraph. Published researcher in Persian NLP, passionate about designing scalable, production-grade AI systems that bridge research and industry.
Experience
Arusha Company · Tehran, Iran
- Core maintainer of Aisa, Arusha's modular suite of Persian NLP microservices (Entity Linking, NER, POS Tagging, Keyphrase Extraction, Summarization, Sentence/Paragraph Splitting, Mention Knowledge Base, Open Information Extraction, LLM Info-Box Extraction); owned service design, model training, and deployment across the suite.
- Designed and built LLM-based agent and tool-use pipelines for autonomous task execution, integrating structured tool-calling with retrieval and reasoning steps.
- Developed and optimized text normalization pipelines using Piraye (PyPI).
- Contributed to knowledge graph design and entity relationship extraction for Persian-language datasets.
- Built hybrid text retrieval systems integrating lexical (BM25) and semantic (vector) methods.
- Integrated LLM-based components into production services for search and question-answering, and worked on LLM serving infrastructure, PostgreSQL, and vector databases for scalable semantic search and RAG.
- Containerized and deployed NLP services using Docker for consistent, scalable production environments.
Ahdsoft Company · Remote
- Worked on NLP models including text similarity, NER, and text summarization.
- Developed HTTP services for entity linking and information extraction.
- Created multilingual text normalizers for English, Persian, and Arabic.
University of Tehran · Tehran, Iran
- Researched and improved Large Language Models for Persian university entrance exam tasks.
- Published at COLING 2025 — “IRUEX: A Study on Large Language Models Problem-Solving Skills in Iran’s University Entrance Exam”.
Selected Projects
Retrieval-Augmented Generation Chatbot
- Implemented a hybrid RAG pipeline combining BM25 and vector-based semantic search.
- Built a RAG dataset and integrated it into a conversational chatbot for factual question answering.
- Employed vLLM, RAGFlow, and Llama.cpp for efficient inference and retrieval.
HodHod Dataset · GitHub
- Created the first large-scale cleaned Persian text dataset with parallelized preprocessing.
- Developed an end-to-end pipeline for tokenization, normalization, and data cleaning.
Piraye Text Normalization Library · GitHub
- Developed a multilingual text normalization library for Persian, Arabic, and English.
- Published on PyPI (piraye) and integrated into enterprise NLP systems.
Persian NER & POS Plugin · GitHub
- Built neural NER and POS tagging models for Persian using deep learning.
- Created a Chrome plugin for real-time Persian text annotation.
Education
University of Tehran · Tehran, Iran
M.Sc. in Artificial Intelligence and Robotics
Shahid Beheshti University · Tehran, Iran
B.Sc. in Computer Engineering
Honors and Certificates
- 3rd Rank in B.Sc. Class, Shahid Beheshti University (2022)
- Top 1% Rank in Iranian University Entrance Exam (2018)
- Deep Learning Specialization — Online Course (2019–2022)
- Data Structures and Algorithms Course — Online (2019)
- Winner — Irancell Labs AI Hackathon, Tehran (2023)
Technical Skills
Programming Languages
Python, Java, C++, MATLAB, R
Machine Learning
PyTorch, TensorFlow, Transformers, Scikit-learn
LLM, Agents & NLP Frameworks
vLLM, RAGFlow, Llama.cpp, LangChain, LangGraph, Hermes, HuggingFace, FastAPI
Agentic AI
Tool-use / function calling, multi-step agent orchestration, agent-tool integration
LLM APIs
OpenAI API, Anthropic API, other LLM provider APIs
Databases
PostgreSQL, SQL
Vector Databases
Semantic search & vector storage for RAG pipelines
NLP Techniques
Text Normalization, Tokenization, Knowledge Graphs, RAG, Semantic Search, Entity Linking, NER, Keyphrase Extraction, Summarization
Data Tools
NumPy, Pandas, SQL, Jupyter, Matplotlib, Power BI
DevOps & Deployment
Git, Docker, Linux
Web Technologies
HTML, CSS, JavaScript, React
Languages
- Persian: Native
- English: Professional Working Proficiency (TOEFL: 100)
- Arabic: Elementary
Interests
- AI Research, Large Language Models, Open-Source NLP, Music, Hiking, Board Games