Skip to content
View Tajaddin's full-sized avatar

Achievements

Achievement: Pull Shark

Achievements

Achievement: Pull Shark

Block or report Tajaddin

Block user

Prevent this user from interacting with your repositories and sending you notifications. Learn more about blocking users.

You must be logged in to block users.

Maximum 250 characters. Please don’t include any personal information such as legal names or email addresses. Markdown is supported. This note will only be visible to you.
Report abuse

Contact GitHub support about this user’s behavior. Learn more about reporting abuse.

Report abuse
Tajaddin/README.md

Tajaddin Gafarov

ML Engineer · LLM Systems · Agentic AI Building production AI systems at scale — LLM-powered agents, RAG pipelines, and backend services that ship.

Tech Stack

LLM & Agents Python LangChain LangGraph HuggingFace PyTorch MLflow pgvector

Backend & Infra FastAPI Node.js TypeScript AWS Docker Kubernetes PostgreSQL Redis

Systems & GPU C++ CUDA

Featured Projects

Project What it does Stack
llm-knowledge-editing LLM evaluation platform: RAG ingestion pipelines, pgvector retrieval, MLflow experiment tracking across ROME, LTE-LoRA, and ICE. Run on 4-GPU cluster over 48 hours. Python, HuggingFace, LLaMA2, LangChain, MLflow
dota2-build-generator Embedding-based recommender trained on 50K+ matches. Deployed as production inference API with TypeScript + Node.js backend, Docker packaging, and CI/CD pipeline. Python, PyTorch, TypeScript, Node.js, Docker
cuda-inlj GPU-accelerated B+ Tree index nested-loop join. 4.2x speedup over CPU baseline on 10M-row datasets, benchmarked within 5% of theoretical bounds. CUDA, C++

Currently Working On

Production LLM agent systems with LangGraph: multi-step tool-use, agentic workflows, and RAG pipelines deployed at scale.

LinkedIn · gafarovtajaddin@gmail.com

Pinned Loading

  1. dota2-build-generator dota2-build-generator Public

    Dota 2 hero build recommender — PyTorch hero/item embeddings, XGBoost matchup classifier, live in-game GSI overlay, OpenDota match data integration

    Python

  2. cuda-inlj cuda-inlj Public

    GPU-accelerated Index Nested-Loop Join using B+ Tree indexing — CUDA C++, parallel B-tree traversal, benchmarked against hash join

    Cuda

  3. llm-knowledge-editing llm-knowledge-editing Public

    Reproducibility study of 'Learning to Edit: Aligning LLMs with Knowledge Editing' (ACL 2024) — benchmarking LTE-LoRA, ROME, and ICE on LLaMA2-Chat-7B across ZsRE and WikiData datasets

    Python