Ayush Goyal (@ayushg1214) 's Twitter Profile
Ayush Goyal

@ayushg1214

Building. ex @microsoft, ex @NimbusEnv (YC W22), @PabioHQ (YC S21)・music junkie・he/him

ID: 3265560020

linkhttps://athenacopilot.ai calendar_today02-07-2015 01:05:41

98 Tweet

102 Followers

835 Following

Ayush Goyal (@ayushg1214) 's Twitter Profile Photo

We’ve been working really hard, and are proud to announce the release of our API and developer documentation! Start here - docs.athenacopilot.ai Athena Copilot 🤖 Enterprise GenAI for teams, customers & partners. 🚀🎊 shoutout to Mintlify - amazing documentation platform!

Ayush Goyal (@ayushg1214) 's Twitter Profile Photo

Been using Cloudflare AI Gateway in production. Works out of the box, and can not imagine not using it in a GenAI project. I get 1. Usage analytics 2. Logs 3. Caching 4. Model fallback 5. Evals all built in one product.

Been using <a href="/Cloudflare/">Cloudflare</a> AI Gateway in production. Works out of the box, and can not imagine not using it in a GenAI project. 

I get

1. Usage analytics 
2. Logs
3. Caching
4. Model fallback
5. Evals

all built in one product.
vLLM (@vllm_project) 's Twitter Profile Photo

After feedback about our v0.8.4 benchmark for DeepSeek R1, we rerun it with suggested changes: vLLM no EP, SGLang updated v0.4.5 -> post1 and EP -> DP, TensorRT-LLM uses overlap scheduler and tuned parameters. We are seeing good results! So why was there a difference? 🧵

After feedback about our v0.8.4 benchmark for <a href="/deepseek_ai/">DeepSeek</a> R1, we rerun it with suggested changes: vLLM no EP, SGLang updated v0.4.5 -&gt; post1 and EP -&gt; DP, TensorRT-LLM uses overlap scheduler and tuned parameters. We are seeing good results! So why was there a difference? 🧵