Auto RecSys speeds up large scale recommendation model experiments
Auto-RecSys: Harnessing Autonomous Research Agents for Industry-Scale Recommender System
Computation and Language
Summary
Recommendation models used by big companies can take days to train, which makes testing new ideas very slow and complicated. The authors introduce Auto-RecSys, a system that runs many experiments at once across multiple servers to speed things up. It keeps track of experiments and learns from past successes and failures, making the process more reliable and reducing the time humans need to spend managing tests. This helps improve recommendation systems more efficiently by automating much of the research process.
What this means in practice
- •For machine learning engineers: Run multiple large-scale recommendation model experiments in parallel with automated error recovery to reduce manual oversight.
- •For cloud infrastructure teams: Support robust distributed execution of long-running machine learning jobs with persistent memory and failure recovery across servers.
Authors
Ming Li, Dai Li, Xuying Ning, Bo Sun, Rui Li, Yi Zhang, Silvia Gong, Xuan Cao, Rui Li, Cornelia Carapcea, Qunshu Zhang, Zhigang Wang, Yinglong Xia, Andy Wang
Abstract
Auto-research agents have shown the potential to automate hypothesis generation, experiment execution, and iterative refinement. However, scaling this paradigm to industry-scale recommendation models introduces two challenges: (1) long feedback loops, where model training can take days, making serial iteration prohibitively slow and requiring parallel exploration across multiple research directions; and (2) system complexity, where large configurations, fragile infrastructure dependencies, and multi-day GPU jobs require robust and recoverable execution. We present Auto-RecSys, an autonomous research system for long-horizon experimentation on industry-scale recommendation models. Auto-RecSys addresses these challenges through three harness designs: (1) distributed asynchronous execution for running multiple experiments in parallel across servers, (2) centralized cross-server memory for persistent and recoverable execution across sessions and failures, and (3) cognitive-procedural separation, where natural-language skill files guide LLM reasoning while deterministic scripts enforce operational correctness. Auto-RecSys further employs a dual-loop self-evolving architecture: an Execution Evolution Loop in which model-specific playbooks accumulate operational knowledge by recording failed attempts and crystallizing successful pipelines, and an Idea Evolution Loop in which experimental outcomes inform subsequent ideation. Evaluated on recommendation models, Auto-RecSys significantly reduces the human time required per experiment cycle and improves execution reliability as its playbooks mature.