Jack Gerrits (@jack_gerrits) profile picture

Software Engineer

New York, NY
Joined September 2019
Photo by Jack Gerrits (@jack_gerrits)
We've been working on this for some time, so I am excited to see this released!
🚀 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. 🔗 github.com/microsoft/magenti…
8
209
Excited to share our #CHI2025 paper: "Interactive Debugging and Steering of Multi-Agent AI Systems"! As 🤖 AI agent teams 🤖 powered by LLMs take on complex tasks, debugging and steering these agents becomes crucial. How can developers better understand AI agent behavior?
2
13
3
55
8,246
Jack Gerrits retweeted
Excited to represent the AutoGen team at this year's Research Forum! Sign up for my talk below! This is joint work with many incredible colleagues especially @ekzhu @jack_gerrits @vykthur @adamfourney @HsseinMzannar @SaleemaAmershi @AhmedHAwadallah @HosnRafa @bigmiao6 and many more people working on @pyautogen who aren't on X: microsoft.com/en-us/research…
AutoGen update! On Feb 25, @bansalg_ will introduce a transformative update that builds on user feedback and redefines modularity, stability, and flexibility to empower the next generation of agentic AI research and applications. Register here: msft.it/6013Umnpl
1
8
26
3,295
Real, deeply integrated async is one of the core features of AutoGen 0.4. It really unlocks many opportunities.
🌻 Python async API + AutoGen AutoGen now uses Python's async API, benefiting agentic workflows by enabling: - Concurrent AI agents and teams - Non-blocking operations (e.g., LLM calls, tool use) - Background tasks Below, e.g., multiple GPT-4o agent teams run concurrently and independently.
1
2
5
1,006
Jack Gerrits retweeted
Working on AutoGen, one of my favorite design principles is: "make it easy for users to do hard things while keeping it easy for users to do easy things." Many agent frameworks introduce complexity in order to handle hard things, but at the same time makes easy things really difficult to do. For example, not everyone needs opinionated abstraction over simple LLM clients and chat history management for building simple agents -- you could and perhaps should do it with just Python. AutoGen v0.4 is different from many frameworks, including its predecessor v0.2: the Core API makes the hard things easy: distributed runtime, multi-tenancy, state management, virtual actors ..., and leave the easy things like calling an LLM API to the users. The event-driven architecture allows you to easily integrate with external systems, and express parallel and asynchronous workflows intuitively. The release of v0.4 paved the way for more innovations and research in the agentic AI space. We are building a foundation and let a thousand flowers bloom.
Announcing AutoGen 0.4, fully reimagined library for building advanced agentic AI systems, developed to improve code quality and robustness. Its asynchronous, event-driven architecture is designed to support dynamic, scalable workflows. Learn more: msft.it/6012ohgli
5
22
2,126
I spent essentially my whole 2024 on AutoGen 0.4 and am so proud of what we produced. This was a huge effort by many, many people. Check out the authors and acknowledgements in the blog post. We have an amazing team!
Announcing AutoGen 0.4, fully reimagined library for building advanced agentic AI systems, developed to improve code quality and robustness. Its asynchronous, event-driven architecture is designed to support dynamic, scalable workflows. Learn more: msft.it/6012ohgli
3
5
23
1,254
Jack Gerrits retweeted
Join us for a talk with @globant and @pyautogen tomorrow!
1
5
553
Jack Gerrits retweeted
It’s great to be part of an amazing team!!
More packed AutoGen sessions at #NeurIPS2024! Thanks to everyone who joined! w/ @ekzhu @jack_gerrits @adamfourney Friederike Niedtner and many others from our team
2
7
1,189
New open-source Python package: markitdown This package makes it easy to convert various files to Markdown (e.g., for indexing, text analysis, etc.). #AutoGen uses it to enable file handling in SOTA agentic systems like #MagenticOne. The API is super simple. cc @adamfourney
2
19
3
81
54,813
Come join us in Discord to discuss the upcoming AutoGen 0.4 and MagenticOne!
We are now also available on Discord!
3
5
526
We are now also available on Discord!
1
3
1
19
2,583
Was great to meet everyone who stopped by!
More packed AutoGen sessions at #NeurIPS2024! Thanks to everyone who joined! w/ @ekzhu @jack_gerrits @adamfourney Friederike Niedtner and many others from our team
1
5
296
Surprise #NeurIPS2024 drop for y'all: phi-4 available open weights and with amazing results!!! Tl;dr: phi-4 is in Llama 3.3-70B category (win some lose some) with 5x fewer parameters, and notably outperforms on pure reasoning like GPQA (56%) and MATH (80%).
18
68
30
404
94,756
Join the new AutoGen Discord community! 🎉 Connect with fellow open source developers and maintainers. 👉 aka.ms/autogen-discord #AutoGen #OpenSourceAgents
2
7
14
1,138
Thanks to everyone who joined the AutoGen workshop at #NeurIPS2024! We will also be at the Microsoft booth on Wed/Thurs: Dec 11: 9:00 AM - 10:00 AM Dec 12: 3:00 PM - 4:00 PM We are hiring strong FTEs (both researchers and engineers) and interns.
5
14
1,068
To our collaborators & community: We’ve seen questions about AutoGen forks/clones vs. the official project. Here's a summary of the latest. Please share with others. - The official repo is github.com/microsoft/autogen. - We're actively working on AutoGen v0.2, with v0.4 innovations like AutoGen-Core, -AgentChat, -Bench, and -Studio 🚀 - Development remains open-source under MIT, and contributions are welcome. - No Microsoft employee is affiliated with AG2, a fork of Microsoft/AutoGen; We recognize varying endeavors/interests in the community and wish them well. ⚠️ Please note: The current pyautogen package isn’t affiliated with Microsoft AutoGen, and admin access is blocked for us. Discord access is also limited for us -- we do not have viewing or posting rights, but you can connect via GitHub Discussions, email ([email protected]), or our Wed office hours (10 AM PST). Thank you for shaping this collaborative project! Stay tuned for more updates soon. 🙌 Read the article below for full details 👇 #AutoGen #OpenSourceAgents #OpenSource
9
24
2
54
24,082
📢As part of our ongoing work to improve AutoGen, we have been rewriting AutoGen Studio. It is still early but here are some of the new features enabled by the AutoGen 0.4 preview API - Real-time agent updates streaming to the frontend - Run control: You can now stop agents mid-execution if they're heading in the wrong direction, adjust the team, and continue - Interactive feedback: Add a UserProxyAgent to get human input through the UI during team runs - Message flow visualization: See how agents communicate with each other Want to participate or ask questions - Join the discussion on GitHub: github.com/microsoft/autogen… #AutoGen #AutoGenStudio #OpenSourceAgents #TechNews
1
10
1
39
2,887
📢Introducing Magentic-One, a generalist 5-agent multi-agent system for solving open-ended web- and file-based tasks. 🤖🤖🤖🤖🤖 Magentic-One represents a significant step towards agents that can complete tasks that people encounter in their daily lives and can achieve strong performance and generalization across THREE challenging agentic benchmarks: GAIA, WebArena, and Assistant. We are releasing an open-source implementation in #AutoGen, our popular open-source framework for developing multi-agent applications. Checkout the technical report, blog, and implementation below 👇 @MSFTResearch @Microsoft #AutoGen #Agents #opensource 1/
3
110
23
410
68,917