🏗️ Building on HF

Dipankar Sarkar PRO

dipankarsarkar

3 200 9

https://www.dipankar.cc

AI & ML interests

Building the AI-native stack. Agents as infrastructure, safety as architecture, performance as plumbing. I publish the receipts: papers, datasets, demos.

Recent Activity

repliedto vineeth98's post about 2 hours ago

I made a speedrun leaderboard for LoRA fine-tuning. One frozen task (Qwen2.5-1.5B to 57% on GSM8K), one GPU, fastest training run wins. Every record gets re-run 3x with fresh seeds on identical hardware before it counts, so no self-reported numbers. The baseline was 11:57 three days ago. Someone already got it down to 1:44, with data pruning and a chunked cross-entropy that never materializes the logits. Attempting is free (Modal's monthly credits cover full runs), and the second track (SmolLM2 + SQuAD) is still sitting at its naive baseline — easy first record for someone. https://huggingface.co/spaces/vineeth98/lora-speedrun

upvoted a paper about 3 hours ago

UI2App: Benchmarking Visual Interaction Inference in Executable Web Application Generation

upvoted a paper about 4 hours ago

WorldCupArena: Fine-Grained Evaluation of Language Models and Deep-Research Agents on Football Forecasting

View all activity

Dipankar Sarkar PRO

AI & ML interests

Recent Activity

Organizations

dipankarsarkar 's collections 2

Before the Pull Request: Mining Multi-Agent Coordination

dipankarsarkar/grite-corpus

Neul Labs — safe agent infrastructure

dipankarsarkar/gpuemu-corpus

The Correctness Illusion in LLM-Generated GPU Kernels

The Correctness Illusion