Gagan Bansal Profile
Gagan Bansal

@bansalg_

Followers
3K
Following
2K
Media
67
Statuses
810

AI research @msftresearch | Built AutoGen @pyautogen | Previously @uwcse, @iitdelhi

Seattle, WA
Joined November 2012
Don't wanna be here? Send us removal request.
@bansalg_
Gagan Bansal
3 months
⚡️How can we build *human-centered* agents?. Building on our research on AutoGen (@pyautogen) and Magentic-One, our team at @MSFTResearch is very excited to release Magentic-UI. This agentic application has several human-centered considerations built into it, improving .-.
@pyautogen
AutoGen
3 months
🚀 Introducing Magentic-UI — an experimental human-centered web agent from @MSFTResearch . It automates your web tasks while keeping you in control 🧠🤝—through co-planning, co-tasking, action guards, and plan learning. 🔓 Fully open-source. We can't wait for you to try it. 🔗
1
2
21
@bansalg_
Gagan Bansal
10 hours
RT @willmcgugan: Hey @googleaidevs @AnthropicAI . I have spent the last 5+ years making things look awesome in the terminal. You could have….
0
139
0
@grok
Grok
2 days
Generate videos in just a few seconds. Try Grok Imagine, free for a limited time.
640
2K
9K
@bansalg_
Gagan Bansal
7 days
RT @JohnCLangford: Thanks :-) . There's a next draft in progress as well:.a) Lesser but still significant (~x1.5 ish) gains in LoRA or FFT….
Tweet card summary image
leloykun.github.io
Simply switching to Muon can already get you 2x efficiency gains. But you can squeeze out an extra 1-2% by optimizing the Newton-Schulz coefficients.
0
2
0
@bansalg_
Gagan Bansal
8 days
RT @rohanpaul_ai: Magentic‑UI shows that putting a person back in charge of plan, pause, and approve steps lets large‑language‑model agents….
0
8
0
@bansalg_
Gagan Bansal
12 days
RT @pyautogen: Interested in the technical details behind Magentic-UI? Our report is now available on arXiv: https:….
0
17
0
@bansalg_
Gagan Bansal
13 days
RT @HsseinMzannar: How do we design and evaluate human-agent interaction? Our report on Magentic-UI details key interaction-patterns includ….
0
9
0
@bansalg_
Gagan Bansal
15 days
No signs
Tweet media one
Tweet media two
0
0
1
@bansalg_
Gagan Bansal
15 days
No matter how advanced AI becomes, one thing will remain essential: its ability to understand human intent. Check out our recent work at @aclmeeting on enhancing "grounding" between people and large language models.
@oshaikh13
Omar Shaikh
16 days
BREAKING NEWS! Most people aren’t prompting models with IMO problems :). They’re prompting with tasks that need more context, like “plz make talk slides.”. In an ACL oral, I’ll cover challenges in human-LM grounding (in 60K+ real interactions) & introduce a benchmark: RIFTS. 🧵
Tweet media one
0
1
10
@bansalg_
Gagan Bansal
21 days
RT @HarshaNori: Incredible work from @thekaransinghal, @rahularoradfs, and the teams @OpenAI and @PendaHealth. Carefully designed real-wo….
0
4
0
@bansalg_
Gagan Bansal
29 days
RT @github: GitHub Copilot coding agent just got a major upgrade. ✨. What's new:.• It tests its own UI changes with Playwright and adds scr….
0
170
0
@bansalg_
Gagan Bansal
30 days
RT @Sumanth_077: Turn any document into LLM-ready data!. Microsoft released MarkItDown a lightweight Python library that converts any docum….
0
145
0
@bansalg_
Gagan Bansal
1 month
RT @ashtom: Being a developer has never been about how fast we can type code. No, being a developer has always been about our ideas. The c….
Tweet card summary image
linkedin.com
The following is a readout of my prepared remarks delivered at WeAreDevelopers World Congress 2025. Hello! I’m Thomas, and I’m a developer.
0
17
0
@bansalg_
Gagan Bansal
1 month
RT @HsseinMzannar: This is really interesting! I'm really surprised by this negative result of AI for coding as it goes against some of my….
0
1
0
@bansalg_
Gagan Bansal
1 month
Most of the new features in the latest version of AutoGen were authored by CoPilot!! @ashtom @ekzhu.
@pyautogen
AutoGen
1 month
🚀 AutoGen v0.6.4 is out!. Shout-out to @github Copilot for helping author these new features!. 🧠 GraphFlow now retains execution state after termination, just like other group chats. Resets only when the graph fully completes. ⚙️ New parameter_override in Workbenches for.
0
3
16
@bansalg_
Gagan Bansal
1 month
RT @simonw: Quitting programming as a career right now because of LLMs would be like quitting carpentry as a career thanks to the invention….
0
1K
0
@bansalg_
Gagan Bansal
1 month
RT @sriniiyer88: Turns out, if you teach llamas how to self-reflect and backtrack from wrong reasoning paths, it does extra well on math re….
0
13
0
@bansalg_
Gagan Bansal
1 month
RT @erichorvitz: Reflections on our recent work on using language models for sequential diagnosis (LinkedIn article). A pleasure collaborat….
0
9
0
@bansalg_
Gagan Bansal
1 month
RT @HarshaNori: As some of you may know, I recently moved to London to help lead a new Health AI team!. Excited for our first research pape….
0
48
0
@bansalg_
Gagan Bansal
2 months
RT @HsseinMzannar: We're doing work to make Magentic-UI more extensible, MCP servers allow you to easily plug in special purpose tools to s….
0
4
0
@bansalg_
Gagan Bansal
2 months
RT @yasik: @gregpr07 @LangChainAI @AnthropicAI One of the best decision we've made early on was to NOT use layers of abstraction and go nat….
0
3
0
@bansalg_
Gagan Bansal
2 months
RT @pyautogen: 🚀 Introducing MCP Agents in Magentic-UI! . Spin up custom agents that wrap one (or many) MCP tools, and let the Orchestrator….
0
27
0