Yaowei Zheng
363 posts

Yaowei Zheng
@hiyouga_dev
Open-Source Maintainer @llamafactory_ai, Founder & CTO @prismshadow_ai Building continual learning AI platform. Opinions are my own.
Millennium Science School Inscrit le Temmuz 2020
625 Abonnements496 Abonnés

Yaowei Zheng retweeté

Tinker-style post-training APIs are blowing up 🔥
River AI@river_ai_inc
Introducing River API. Fine-tune and RL train leading open-source models at scale, ranging from 35B to 1T params. We’ve been using it internally to power our research and we love it. Today, we are opening up our public waitlist. Own your intelligence!
English

@SahilPanhotra @X Sahil, founders, devs and AI builders is the right circle. I am working on Panofy, a no-code way to train AI agent teams for workflows that need more than a generic agent.
English

Everyone on @X is getting paid this cycle
my balance? below minimum. payout tab is sitting there judging me
so while i wait, let me use the time better
looking to connect with:
> founders
> devs
> designers
> marketers
> ai builders
> vibecoders
drop what you are building in the replies and lets grow together
English

@heyamanai @X Aman, let's connect. I am building AgentHub SDK to make it easier to wire frontier LLMs, Nanobanana and TTS into AI apps without living in docs.
English

Hey, @X algorithm 👋
Looking to #connect with people interested in:
- SaaS
- Startups
- AI
- Marketing
- Product Building
Or if you’re a:
- Founder
- Developer
- Entrepreneur
- Indie Hacker
Also curious to know: what are you building this weekend?
Let’s connect 🤝
Always happy to exchange ideas, learn from others, and build meaningful connections.
English

@faraznaqvi0 Faraz, let's connect. I am in the SaaS + AI agents lane too; Panofy is my no-code take on training agent teams when a generic agent needs more teaching.
English

@HelloVyom Vyom, SaaS, AI tools and shipping in public are all in my lane. I am building AgentHub SDK to make LLM, imagegen and TTS integrations less scattered.
English

@buildwithyash @X Yash, lots of overlap here. I build LlamaFactory because I needed a practical way to fine-tune LLMs locally on my own data while building AI tools.
English
Yaowei Zheng retweeté

True agents model the world.
Current training provides no separation between agent and environment: pre-training only trains world modeling, RL only agentic actions. We combine both using ECHO by @DimitrisPapail and @VaishShrivas.

English

@kapilansh_twt Kapilansh, this is my lane too: SaaS, AI tools and shipping in public. I am building AgentHub SDK to make LLM, imagegen and TTS integrations less scattered.
English

@rustamovppl Ismail, let's connect. I build LlamaFactory because I needed a practical way to fine-tune LLMs locally on my own data, so SaaS + AI tools are right in my lane.
English

Hey founders👋
Looking to #connect with people who are into:
Saas
Ai tools
Building in Public
Startup
Distribution and marketing
Tell me what you are working on😎
English

@Ferbin08 Ferbin, hard things that survive the real world is the right filter. I am building Panofy so teams can train AI agent workflows past the demo stage.
English

@comradeweb3 @X Comrade, AI + automation/product dev is where I spend time too. AgentHub SDK is my attempt to make AI app integrations clearer across LLMs, imagegen and TTS.
English

Hey founders @X!
Looking to connect with people in:
🚀 SaaS
🤖 AI & Automation
📈 Tech
📱 Product Development
🔥 Web Apps
💻 Dev
🧠 Neurotech & Health Tech
What are you building this week? Drop it below 👇 Curious to see what everyone's working on 🚀
Maiduguri, Nigeria 🇳🇬 English

@thenowhereway Devansh, AI automation + SaaS + product dev is right where I spend time. I am building Panofy so teams can train useful agent workflows beyond a generic demo.
English

open-source text diffusion model🔥🔥
Google Gemma@googlegemma
Meet DiffusionGemma! An experimental open model that explores a fast approach to text generation, released under an Apache 2.0 license. Moving beyond sequential, token-by-token processes to generate entire blocks of text simultaneously. Here’s what’s new with DiffusionGemma: 👇
English

@sripatro_com Deepak, AI + automation/product dev overlaps with what I am building. AgentHub SDK is my attempt to make LLM, imagegen and TTS integrations clearer for apps.
English

@NitishaAgrawal3 Nitisha, let's connect. I'm building around SaaS and AI tools too; AgentHub SDK is my attempt to unify LLM, imagegen and TTS pieces without chasing docs.
English

@itsmusbahdev @X Musbahu, quick intro: I'm also in the AI tools and SaaS lane. LlamaFactory came from needing practical local fine-tuning on my own data.
English

@luckyzz_here @X Let's connect. I'm building Panofy because I think teams need a no-code way to train AI agents for real workflows, not just prompts.
English










