Search papers, labs, and topics across Lattice.
Deep Researcher Agent is an open-source framework enabling LLM agents to autonomously manage the entire deep learning experiment lifecycle, from hypothesis to refinement. It introduces zero-cost monitoring via process-level checks, a constant-size memory architecture to prevent unbounded context growth, and a minimal-toolset leader-worker architecture to reduce token overhead. Deployed for 30+ days, the framework completed 500+ experiment cycles across four projects, achieving a 52% improvement in one project at a cost of only $0.08 per 24-hour cycle.
Run 500+ deep learning experiments for just $0.08 per day with this new autonomous agent framework.
We present \textbf{Deep Researcher Agent}, an open-source framework that enables large language model (LLM) agents to autonomously conduct deep learning experiments around the clock. Unlike existing AI research assistants that focus on paper writing or code generation, our system addresses the full experiment lifecycle: hypothesis formation, code implementation, training execution, result analysis, and iterative refinement. The framework introduces three key innovations: (1) \textbf{Zero-Cost Monitoring} -- a monitoring paradigm that incurs zero LLM API costs during model training by relying solely on process-level checks and log file reads; (2) \textbf{Two-Tier Constant-Size Memory} -- a memory architecture capped at $\sim$5K characters regardless of runtime duration, preventing the unbounded context growth that plagues long-running agents; and (3) \textbf{Minimal-Toolset Leader-Worker Architecture} -- a multi-agent design where each worker agent is equipped with only 3--5 tools, reducing per-call token overhead by up to 73\%. In sustained deployments spanning 30+ days, the framework autonomously completed 500+ experiment cycles across four concurrent research projects, achieving a 52\% improvement over baseline metrics in one project through 200+ automated experiments -- all at an average LLM cost of \$0.08 per 24-hour cycle. Code is available at https://github.com/Xiangyue-Zhang/auto-deep-researcher-24x7.
