A journey through intelligence
Mohammad Javad
Ghaderi poor
AI Engineer / LLM & RAG Engineer
Enter the systemMohammad Javad
The input layer
Context changes
everything.
I build context-aware AI systems with large language models, semantic retrieval, and external knowledge. My work connects end-to-end RAG engineering, agentic workflows, and model evaluation with research in neural network interpretability. I am completing an M.Sc. in Artificial Intelligence at Allameh Tabataba'i University.
M.Sc. in Artificial Intelligence
Allameh Tabataba'i University · 2024 - Present
B.Sc. in Computer Science
Shahrekord University · 2020 - 2024
Tokenization
A language
for building.
Language becomes tokens. Tokens become representations. These are the tools I use to build the system around them.
LLM frameworks & tools
LangChain / ChromaDB / Hugging Face / OpenAI API / Anthropic Claude API / Google Gemini API / Ollama / Gradio
Programming & development
Python / SQL / Git & GitHub / Jupyter Notebook / REST APIs / API Integration / Object-Oriented Programming / Android Application Development / UI/UX Design
Embedding space
Meaning has
a position.
Related concepts gather into neighborhoods. A query becomes a vector, so retrieval can follow meaning beyond exact words.
A spatial interpretation of semantic similarity.
My focus
Large Language Models / Retrieval-Augmented Generation / Semantic retrieval & vector embeddings / AI agents & multi-agent systems
Retrieval foundation
Semantic Search / Information Retrieval / Vector Embeddings / ChromaDB
Retrieval-augmented generation
Intelligence
needs context.
Context-aware applications that connect language models with external knowledge through semantic retrieval.
Vector embeddings → vector database → semantic retrieval → LLM-based generation; integrated with tool calling, agents, and local inference.
Encode the queryTOP K = 3
Illustrative vectors. Cosine-ranked retrieval. Scroll to assemble the context.
Python / LangChain / Hugging Face / ChromaDB / Ollama / LLM APIs
Inside attention
Learning
what matters.
Queries meet keys. Attention weights select how values contribute to each token’s next representation.
Attention(Q, K, V) = softmax(QKᵀ / √dₖ)V
Three illustrative heads; twelve tokens. The highlighted query changes as you travel.
My practice
LoRA / QLoRA / Parameter-Efficient Fine-Tuning / Local Inference / LLM Evaluation / Open-Source Language Models
Develop frontier and open-source LLM applications with tool calling, AI agents, multi-agent workflows, and local inference.
Professional experience
09/2025 - Present · Tehran, Iran
LLM & RAG Researcher / AI Engineer
Researching and developing LLM, RAG, generative AI, and agentic systems for context-aware applications.
- Build end-to-end RAG pipelines with embeddings, vector databases, semantic retrieval, and LLM-based generation.
- Develop frontier and open-source LLM applications with tool calling, AI agents, multi-agent workflows, and local inference.
- Apply LoRA/QLoRA, parameter-efficient fine-tuning, and LLM evaluation.
Python / LangChain / Hugging Face / ChromaDB / Ollama / LLM APIs / LoRA / QLoRA
Experience layer 1 of 5
Applied intelligence
RAG systems
Work drawn from LLM & RAG experience
Context-aware applications that connect language models with external knowledge through semantic retrieval.
- Architecture
- Vector embeddings → vector database → semantic retrieval → LLM-based generation; integrated with tool calling, agents, and local inference.
- Contribution
- Research and development of end-to-end retrieval pipelines and frontier or open-source LLM applications, including model evaluation and parameter-efficient fine-tuning.
Python / LangChain / Hugging Face / ChromaDB / Ollama / LLM APIs
Work module 1 of 3
Generation
Research becomes
contribution.
Context informs intelligence.
AdaLayer-CAM: Adaptive Multi-Layer Method for Visualization and Explanation of CNNs ↗IEEE · 2026
2026Ph.D. entrance exam · Rank 19
2025Top Master's Student
2023M.Sc. entrance exam · Rank 190
2023Top Undergraduate Student
Learning & credentials +
- LLM Engineering: Master AI and Large Language Models · Udemy
- Supervised Machine Learning ↗ · Coursera
- Advanced Learning Algorithms
- Unsupervised Learning, Recommenders, Reinforcement Learning ↗ · Coursera
- Digital Marketing ↗ · Google
English B2
The output is a beginning
Let’s build
with context.
Language models. External knowledge.
Intelligent systems that connect them.
mohammad.ghaderi211@gmail.com
Download resume ↓Return to input ↑