AI Engineer World's Fair 2026 Vision & OCR Track Goes Live
The track's framing matters: current VLMs fail at basic visual grounding tasks like counting 32 squares on a chessboard, which explains why document and OCR pipelines still need engineered context layers. Builders working on RAG or agentic document workflows can pick up concrete patterns from LlamaIndex and Reducto sessions on ingestion and context design. Sarvam, an India-based lab, also shows how a 3B state-space vision model is trained outside the usual frontier-lab playbook.
AI Engineer has released the Vision & OCR Track recordings from AI Engineer World's Fair 2026 on YouTube.
The track includes eight talks, with speakers from LlamaIndex, Hugging Face, Meta, Reducto, Elorian, Firecrawl, Perceptron AI and Sarvam.
Highlights include Jerry Liu's session on building a document context layer for AI agents and Merve Noyan's talk on using vision language models with Skills instead of deploying them standalone.
Sarvam's talk covers training a 3B state-space vision model from scratch, and the organisers summarise the track's thesis as: the models can see, they are still learning to look.
Live now: our Vision & OCR Track from AI Engineer World's Fair 2026. A model that counts 32 white squares on part of a chessboard. A file format that stores a table as a pile of line segments. Ten turkeys on the roof of a Tesla. Thesis: the models can see. They are still learning to look. youtube.com/watch?v=RQi7x-… - Building the Document Context Layer for AI Agents: @jerryjliu0 , LlamaIndex - Skill issue: stop deploying vision language models, use them with Skills: @mervenoyann , Hugging Face - Modality Misalignment and Originality Attribution in Short-Form Video: Aditya Gautam, Meta - From Ingestion to Agents: How AI Teams Build on Document Intelligence: Adit Abraham, Reducto - The Best Models Still Reason Like Toddlers: @andrewdai , Elorian - You're Not Thinking Big Enough: Rebuilding Food Systems with AI Agents: @cbmenefee , Firecrawl - From VLM/VLA's to Embodied Agents: @ArmenAgha , Perceptron AI - From Scratch to SOTA: Training a 3B State-Space Vision Model: @fewshotlearner , Sarvam📊 4 ⚡
This report is categorized under AITOP's "Industry" desk, focusing on US frontier AI labs and Silicon Valley technology leaders. Within our cross-regional analytical framework, it directly maps to AI developer toolchains, coding agents, and engineering workflows. Key entities and technical topics under continuous tracking include LlamaIndex, Hugging Face, OCR, vision-language models.
Source attribution: curated from AI Engineer (@aiDotEngineer). Cross-source corroboration: 3 independent sources have covered this development (including The Information, Jerry Liu), confirming industry-wide signal strength.
