Lightnews — Scholar-powered news

Aslan

@aslandizaji.bsky.social

330 followers 1.3K following 65 posts

Artificial Intelligence, Machine Learning, Neuroscience, Complex Systems, Economics. PhD Student at the University of Tehran. Cofounder: @AutocurriculaLab, @NeuroAILab, @LangTechAI. https://sites.google.com/a/umich.edu/aslansdizaji/

sites.google.com

Posts Media Videos Starter Packs

Pinned

Aslan @aslandizaji.bsky.social · Nov 30

Here is the code of the project:
github.com/aslansd/modi...
github.com/aslansd/modi...

Here is the paper of the project:
arxiv.org/abs/2411.17724

Incentives to Build Houses, Trade Houses, or Trade House Building Skills in Simulated Worlds under Various Governing Systems or Institutions: Comparing Multi-agent Reinforcement Learning to Generative...

It has been shown that social institutions impact human motivations to produce different behaviours, such as amount of working or specialisation in labor. With advancement in artificial intelligence (...

arxiv.org

2 1 6

Aslan @aslandizaji.bsky.social · 4d

Finally, this app was inspired by two courses from @DeepLearningAI and @LangChainAI Academy. I would like to thank them!

Aslan @aslandizaji.bsky.social · 4d

Check it out here 👇
GitHub (code): github.com/aslansd/deep...
Streamlit app (try it out): d3yjnms2ch6yxcthmrvqnn.streamlit.app
Would love feedback from researchers, builders, and anyone interested in multi-agent systems 🚀

GitHub - aslansd/deepagent_researcher_analyse_visualise: Build a deep agent researcher which analyses and visualises the information

Build a deep agent researcher which analyses and visualises the information - aslansd/deepagent_researcher_analyse_visualise

github.com

1 3

Aslan @aslandizaji.bsky.social · 4d

The results?
📊 Final answers with explanations
📈 Automatically generated charts
📝 Full trace of every step the system took
It’s like having a research assistant that works systematically and shows its work!

1 1

Aslan @aslandizaji.bsky.social · 4d

My app combines all of this into a multi-agent framework:
A Planner breaks down your research question
Specialized agents (web researcher, chart generator, summarizer) handle subtasks
Everything is orchestrated into a transparent workflow you can trace

1 1

Aslan @aslandizaji.bsky.social · 4d

Traditional LLM agents can be “shallow” — they just loop through tools and struggle with long, complex tasks.
Deep Agents (a new feature in LangGraph) bring:
✅ Task planning (TODOs)
✅ Sub-agent delegation
✅ Context offloading to files
✅ Robust reasoning prompts

1 1

Aslan @aslandizaji.bsky.social · 4d

It’s a Streamlit app powered by LangGraph Deep Agents that can take your research question, plan a workflow, fetch data, generate charts, and explain results step by step.

1 1

Aslan @aslandizaji.bsky.social · 4d

🚀 Excited to share something I’ve been building: the Multi-Agent Deep Research Assistant 🧠✨

🔗 GitHub: github.com/aslansd/deep...
🔗 App: d3yjnms2ch6yxcthmrvqnn.streamlit.app

GitHub - aslansd/deepagent_researcher_analyse_visualise: Build a deep agent researcher which analyses and visualises the information

Build a deep agent researcher which analyses and visualises the information - aslansd/deepagent_researcher_analyse_visualise

github.com

1 2

Aslan @aslandizaji.bsky.social · 23d

Big thanks to @DeepLearningAI & @LangChainAI Academy for the resources that made this possible.

Aslan @aslandizaji.bsky.social · 23d

Multi-Modal RAG App (Streamlit + Ollama)
Built a lightweight Retrieval-Augmented Generation (RAG) system that processes both text + image docs. Users can load files, build a vector store, and run retrieval-grounded QA.
github.com/aslansd/mult...

GitHub - aslansd/multi-modal-rag-web: A Multi-Modal RAG Application Built with Streamlit Using a Lightweight Ollama Model

A Multi-Modal RAG Application Built with Streamlit Using a Lightweight Ollama Model - aslansd/multi-modal-rag-web

github.com

Aslan @aslandizaji.bsky.social · 23d

React Native + Expo Go App for Open Deep Research
Brought the same framework to mobile, enabling research automation on the go.
github.com/aslansd/open... expo.dev/accounts/asa...

GitHub - aslansd/open_deep_reasearch_mobile: A React Native plus Expo Go App for open_deep_research Framework of LangChain/LangGraph

A React Native plus Expo Go App for open_deep_research Framework of LangChain/LangGraph - aslansd/open_deep_reasearch_mobile

github.com

Aslan @aslandizaji.bsky.social · 23d

Streamlit App for Open Deep Research
Adapted LangChain’s Open Deep Research (LangGraph) into a Streamlit app to support advanced research workflows: ingestion, retrieval, multi-modal analysis, and context-aware Q&A.
github.com/aslansd/open...
rneaovknvzddyykhkeu2et.streamlit.app

GitHub - aslansd/open_deep_research_web: A Streamlit App for open_deep_research Framework of LangChain/LangGraph

A Streamlit App for open_deep_research Framework of LangChain/LangGraph - aslansd/open_deep_research_web

github.com

Aslan @aslandizaji.bsky.social · 23d

Over the past two weeks at LangTechAI, I built 3 generative AI apps on top of LangChain / LangGraph frameworks — inspired by courses from @DeepLearningAI & @LangChainAI Academy.

1 1