OPEN-SOURCE FRAMEWORK
A-Evolve
A shared framework for building evolvers that turn agent experience into durable improvements. The foundation for our work on online and adaptive harness learning.
Recursive self-improvement · Autonomous research
Hanqing (Henry) Lu
Founder & Research Lead, A-EVO Lab at Amazon
I study how AI can improve AI—from evolving agent harnesses to building autonomous research systems, evaluating their discoveries, and training models for research.
I lead A-EVO Lab at Amazon and co-lead OpenRSI Index. Previously, I worked at Microsoft Research and studied machine learning at Carnegie Mellon University (CMU).
A research lab at Amazon focused on recursive self-improvement.
Explore the lab ↗Research
01
RSI-Harness
Learning from experience through persistent improvements to skills, memory, prompts, and workflows.
EMNLP 2026 · MAIN
Evo-Harness: Context-to-Harness Skill Compilation for Self-Evolving Agents
Compiling noisy, one-shot execution experience into reusable skills for online harness learning.
EMNLP 2026
Harness Updating Is Not Harness Benefit: Disentangling Evolution Capabilities in Self-Evolving LLM Agents
Separating the ability to produce harness updates from the ability to benefit from them.
EMNLP 2026 · MAIN
MemMA: Coordinating the Memory Cycle through Multi-Agent Reasoning and In-Situ Self-Evolution
Coordinating memory construction and retrieval, with verification and repair before memory is finalized.
arXiv · 2026
Adaptive Auto-Harness: Sustained Self-Improvement for Agentic System Deployment on Open-Ended Task Streams
Sustaining agent performance on changing task streams through a stateful evolver, a harness tree, and task-wise routing.
02
Autoresearch Harness
Systems that organize and execute the research process: hypotheses, experiments, evaluation, and revision.
arXiv · 2026
A-Evolve-Training: Autonomous Post-Training of a 30B Model
An autonomous system conducts a multi-week post-training campaign and revises its search policy when its internal proxy becomes misleading.
Reported score: 0.86 · top human submission: 0.87
03
RSI-Evaluation
Evaluating autonomous research against real model-development problems and human baselines.
BENCHMARK · PROJECT CO-LEAD
OpenRSI Index
An open benchmark for AI research agents, built around auditable tasks and fixed verifiers.
My contribution: co-led the project; designed, implemented, and ran the benchmark, including its Signature Tasks.
Marin-Scaling-Ladder · Qwen-122B-RL-Merge · GPIC
04
RSI-Training
Training models for autonomous research.
Autoresearch Models
A forthcoming research direction.
05
Agents & Reinforcement Learning
Selected work on agent training, tool-use evaluation, and value-guided decoding.
ACL 2026
ICLR 2026
NeurIPS 2017