⏺ Read(about.md)
⎿ Read 3 lines
AI Research Engineer with a deep passion for neural networks, large language models, and scalable ML systems. Strong foundation in PyTorch and CUDA; committed to open-source development and to bridging the gap between research and real-world applications by translating papers into practical implementations.
see also: portfolio ↗
⏺ Read(publications/)
⎿ 7 papers · ordered by most proud of
- ExpertRAG: Efficient RAG with Mixture of Experts — Optimizing Context Retrieval for Adaptive LLM Responses
- Galvatron: Automatic Distributed Training for Large Transformer Models
- Theoretical Foundations and Mitigation of Hallucination in Large Language Models
- Mixture of Transformers: Macro-Level Gating for Sparse Activation in Large Language Model Ensembles
- Bachelor Thesis: AI Engine: Deep Learning and Neural Network Engine
- Universal Approximation Theorem for a Single-Layer Transformer
- Mixture of Attention Schemes (MoAS): Learning to Route Between MHA, GQA, and MQA
⏺ Bash(ls ~/projects)
⎿ 6 directories
- nanograd — ML/DL ecosystem engine: GPT, Llama, Stable Diffusion, ViT ·
Python - Axon — paper implementations: InstructGPT, Llama, transformers, diffusion ·
Python - Nexus — dynamic GPU allocation for Mixture-of-Experts ·
Python - NeuroFlow — node-based AI training pipelines ·
Python - Galvatron — large-scale distributed transformer training ·
Python - ExpertRAG — MoE-routed retrieval-augmented generation ·
Python
⏺ Bash(esmail --status)
Core Python · PyTorch · CUDA
Research LLMs · Mixture-of-Experts · RAG · diffusion models
Systems distributed training · GPU scheduling · inference optimization
Tools Hugging Face · Git · Linux
⏺ Bash(gh stats --user Esmail-ibraheem)
⎿ stats: snapshot Jul 2026 · streak: refreshed daily
⏺ Read(contact.md)
Portfolio · LinkedIn · Google Scholar · Hugging Face · ORCID · ResearchGate · Semantic Scholar
~/esmail-ibraheem · esmail-v25 · ✻ 128k context · ? for shortcuts