Open to opportunities

mumong wang

@mumongwang

Cloud-native LLM application engineer focused on reliable agent delivery, context optimization, and evaluation-driven iteration.

China

Message

What I'm looking for

I’m looking for a team where I can build reliable LLM agent workflows end-to-end—context optimization, observability, tracing, and evaluation loops—while keeping real-world execution deterministic. I want to grow in cloud-native AI infra and production stability.

I’m a cloud-native LLM application/agent engineer who focuses on making agent systems stable, traceable, and reliably iterative. I improve Agent stability and iteration through task orchestration, context management, execution tracing, and evaluation loops—especially under small-model constraints.

I build agent workflows that understand capability boundaries and apply deterministic logic when real-world tasks shouldn’t be delegated to LLMs. My core strengths include workflow orchestration, MCP tool use, structured handoff, output constraints, prompt caching, tool-output compression, state summaries, and fallback logic.

On the platform side, I design and deliver observability data foundations for AIOps and agent diagnosis, integrating Kubernetes-native tooling and AI inference deployment stacks. I’ve researched and deployed AI infra tools such as KubeFlow, vLLM/LMDeploy, and Volcano/Ray, and I productize reliable deployment and evaluation with clear tracing and abnormal-case replay.

Experience

Work history, roles, and key accomplishments

Current

Cloud Platform R&D Engineer

Current

Zhongguancun Panlink

Jun 2024 - Present (2 years)

Built a four-node AIOps Agent diagnosis workflow and a fast query path for stable, traceable execution under small-model constraints, focusing on context optimization and evaluation. Researched and deployed AI infrastructure tooling (KubeFlow, vLLM/LMDeploy, Ray/Volcano) and used AI coding QA practices to formalize SDD, code review, and bad-case review into the development workflow.

Workflow Orchestration Agent Evaluation Loops Kubernetes Prometheus EFK

DevOps Intern

Valtech China

Oct 2023 - Jan 2024 (3 months)

Maintained CI pipelines for connected-vehicle services by diagnosing build failures, service issues, and environment configuration problems. Configured Nginx mutual TLS and a containerized test proxy to support testing workflows.

Build Failure Diagnostics Nginx

Cloud Computing Dev Intern

Transwarp

Jul 2022 - Oct 2023 (1 year 3 months)

Contributed to cloud platform product deployment development through automation scripts and multi-OS adaptation. Supported Kubernetes troubleshooting and provided technical support during delivery and operations.

cloud deployment Technical Support