Dean @codevore1

building opensource Document Intelligence @ docuglean-ai. Documents → structured data https://t.co/kygHLpauf6 RLHF. RL Envs. git.new/docuglean-ocr Joined June 2018

Tweets

320
Followers

303
Following

462
Likes

153

Dean @codevore1

4 days ago

Excited to announce I’m building Docuglean - an open-source, privacy-focused document intelligence layer with three key foundations:. • Agentic OCR for parsing documents into high-quality, ready-to-use data. powered by SOTA VLM models for complex tables, formats, and layouts.…

1 1 21 388 0

Download Image

Dean @codevore1

a week ago

I reviewed 100 models over the past 30 days. Here are 5 things I learnt. TL;DR: Spent a month testing every AI model for work, a few tools I'm building and RL. Build task-specific evals. Most are overhyped, a few are gems, model moats are ephemeral, and routers/gateways are the…

0 1 1 55 0

Dean @codevore1

3 weeks ago

High-quality data is the fuel for modern data deep learning model training. HT @lilianweng lilianweng.github.io/posts/2024-02-…

0 0 0 66 0

Dean @codevore1

4 weeks ago

Human evaluation still reigns for subjective cases: edge-case nuance escapes auto-scoring

0 0 0 35 0

Dean @codevore1

4 weeks ago

always love seeing tiny teams build really impressive models. know any other small team doing cool model work?

Divam Gupta @divamgupta

a month ago

always love seeing tiny teams build really impressive models. know any other small team doing cool model work?

179 544 5K 338K 4K

Download Video

0 0 0 69 0

Dean @codevore1

4 weeks ago

you cant outcode your way out of marketing

0 0 0 27 0

Dean @codevore1

a month ago

@Alibaba_Qwen silently dropping a SOTA image model with amazing txt rendering that beats GPT Image 1 and Seedream 🔥 def wasn't on my wish list this week! 🤯