EnCompass: Enhancing Agent Programming with Search Over Program Execution Paths

Zhening Li, Armando Solar-Lezama, Yisong Yue, Stephan Zheng

Keywords: AI agents, large language models, agent frameworks, angelic nondeterminism, inference-time strategies, test-time scaling

TL;DR: programming framework that does runtime compilation of an LLM-based agent workflow into a search space, enabling independent experimentation of different overlaying inference-time search strategies

Abstract: We introduce a new approach to *agent programming*, the development of LLM-based agents. Current approaches to agent programming often entangle two aspects of agent design: the core workflow logic and the inference-time strategy (e.g., tree search). We introduce *probabilistic angelic nondeterminism* (PAN), a programming model that disentangles these two concerns, allowing the programmer to describe the agent workflow and independently experiment with different inference-time strategies by simply changing a few inputs. We provide an implementation of PAN in Python as the EnCompass framework, which uses a Python decorator to compile agent workflow programs into a search space. We present three case studies that demonstrate how the framework lets the programmer quickly improve the reliability of an agent and easily switch between different inference-time strategies, all with little additional coding.

Primary Area: Infrastructure (e.g., libraries, improved implementation and scalability, distributed solutions)

Submission Number: 11511