Dinkar Juyal

I am an ML Engineer at Meta, where I’m the tech lead for agent quality on Meta’s Business AI Agents — the feedback loops, evals, training data synthesis, and agentic harness work behind an agent now facilitating 10M+ conversations a week across 200M+ businesses. Previously, I was a Staff ML Researcher and Engineer at PathAI, focused on foundation models and computer vision for digital pathology.

My recent work is on credit assignment and evaluation for long-horizon agents: Self-Credit (dense supervision from sparse outcome rewards via intra-group rescaling), CDBench (controlled multi-fault debugging with RL and iterative repair), and ML-Dev-Bench (evaluating AI agents as ML engineers). Earlier work spans foundation models for pathology (PLUTO), domain generalization (ContriMix, best-performing official submission on the Camelyon17 WILDS leaderboard), and interpretability (Additive MIL), applied clinically in AI-based automation of enrollment criteria and endpoint assessment in liver disease trials (Nature Medicine, 2024). See publications for the full list.

I write about ML research — mostly notes on agents, foundation models, and evaluation. New posts go up on Substack first (see below); some are later mirrored here.

Talks & articles

• PyTorch Conference 2023 — Lightning talk on domain generalization in medical imaging
• ML Seminar Series, University of Minnesota — AI for digital pathology