Pratik Karki @ai_evals

co-founder at @anthromindinc | ex-google AI engineer | building next-gen scalable oversight systems for AI | https://t.co/2EFb7GMpjC anthromind.com San Francisco Joined May 2025

Tweets

194
Followers

90
Following

371
Likes

204

Pratik Karki @ai_evals

14 hours ago

TL;DR: @thinkymachines tinker good. Managed fine-tuning good. Need humans in loop or else enterprise AI no work. Thinking Machines dropped Tinker yesterday, and it's actually pretty clever. I hope to "tinker" with it as soon as I get access :) Most companies I speak with are…

0 0 3 55 0

Download Image

Pratik Karki @ai_evals

16 hours ago

Got a great customer use case on the paraphrasing problem for their data catalog. I was just in a webinar with @askalphaxiv, and the presenter literally brought up new research on the topic before my customer call. Life can be serendipitous :)

0 0 3 25 0

Pratik Karki @ai_evals

19 hours ago

Happy #Dashain2025 to all my Nepali folks! Especially if you're in the US and working this week, like I am 🥲. Enjoy the festivities and time with your family. Well, I'm off to my next customer meeting.

0 0 1 24 0

Pratik Karki @ai_evals

3 days ago

Stumbled into @ShashwatGoel7's living substack doc on the path to AGI. Quite an amazing resource! The scalable oversight mission of @AnthromindInc works well with the thoughts presented here. Although AI agents do well in olympiads and whatever, the true test is how well they'll…

0 0 1 53 0

Download Image

Andrew Ng @AndrewYNg

3 days ago

Announcing a significant upgrade to Agentic Document Extraction! LandingAI's new DPT (Document Pre-trained Transformer) accurately extracts even from complex docs. For example, from large, complex tables, which is important for many finance and healthcare applications. And a…

90 572 4K 271K 4K

Download Video

Pratik Karki @ai_evals

3 days ago

Thousands of UGC startups cried out and disappeared overnight

OpenAI @OpenAI

3 days ago

Thousands of UGC startups cried out and disappeared overnight

3K 4K 45K 19.9M 15K

Download Video

0 0 1 62 0

ℏεsam @Hesamation

3 days ago

Stanford’s CS336 "Language Modeling from Scratch" is available for free on YouTube. Instead of wasting time on background, it goes straight into practical topics you rarely see explained elsewhere — PyTorch, MoE, Triton, parallelism, eval, scaling laws, alignment, and more.

7 59 442 15K 408

Download Image

ry @rywalker

4 days ago

Human-in-the-loop is critical capability for real-world AI implementations, love that Airflow can help teams move beyond POCs and into careful real-world usage.

Astronomer @astronomerio

4 days ago

Human-in-the-loop is critical capability for real-world AI implementations, love that Airflow can help teams move beyond POCs and into careful real-world usage.