ebarkhordar

Follow

☄️

Ehsan ebarkhordar

☄️

Follow

NLP researcher: bias detection, multilingual NLP, the social-science side of LLMs. Now mostly agentic AI and how agents break. Ex-backend dev.

64 followers · 179 following

Achievements

Achievements

Highlights

Developer Program Member

ebarkhordar/README.md

Hi there, I'm Ehsan Barkhordar 👋

🛠️ Tech Stack

🔧 Recent Upstream Work

changedetection.io #4249: made the restock notification token safe outside restock watches (merged)
local-deep-research #5062: FAISS cosine scores were inverting the search ranking (in review)
swarms #1712: stop sending empty text blocks to Anthropic on image-only vision calls (in review)
llama_index #22333, #22337: LLM adapter fixes for Cohere and SambaNova (in review)
camel #4170: Hermes ShareGPT round-trip was dropping tool-call arguments (in review)

📊 GitHub Stats

Pinned Loading

political-inclination political-inclination Public archive

This repository implements a multilingual political inclination analysis tool, leveraging advanced natural language processing techniques to detect political and economic biases in textual data for…

Jupyter Notebook
voter-behavior-prediction-LLM voter-behavior-prediction-LLM Public archive

This project explores the predictive power of large language models (LLMs) on voter behavior using social media data. By integrating demographic information with user-generated content such as twee…

Jupyter Notebook 5
shilling-attack-detection-for-recommender-systems shilling-attack-detection-for-recommender-systems Public archive

shilling attack detection for recommender systems

Jupyter Notebook 7 2
GGLAB-KU/LOBSTER GGLAB-KU/LOBSTER Public

LOBSTER: A dataset and code for analyzing language bias in academic peer reviews

Python 1