Zhilin Xu said: Here’s the same intro with no em dashes, written in a natural, client-ready voice: My name is Zhilin Xu, and I’m a machine learning engineer who focuses on building AI systems that actually run in production. Over the last several years I’ve designed and shipped large-scale ML and LLM solutions for enterprise products. That includes real-time diagnostic intelligence at AppDynamics, multi-agent conversational systems for education, and high-performance model deployment using Python, PyTorch, FastAPI, and cloud tools. Before that, I worked at Black Sesame on automotive AI toolchains, where I built graph optimization and quantization modules and deployed perception models onto embedded hardware for OEMs. I’ve spent a lot of time owning full pipelines from data processing and model training to serving, monitoring, and frontend integration. What makes me stand out is that I treat AI as an engineering problem, not a research exercise. I focus on reliability, low latency, and clean design, and I’m used to taking complex workflows and turning them into something fast, stable, and easy for users to depend on. I enjoy solving real problems with technology, and I’m comfortable taking responsibility for systems end to end.

Zhilin Xu

Zhilin Xu said: Here’s the same intro with no em dashes, written in a natural, client-ready voice: My name is Zhilin Xu, and I’m a machine learning engineer who focuses on building AI systems that actually run in production. Over the last several years I’ve designed and shipped large-scale ML and LLM solutions for enterprise products. That includes real-time diagnostic intelligence at AppDynamics, multi-agent conversational systems for education, and high-performance model deployment using Python, PyTorch, FastAPI, and cloud tools. Before that, I worked at Black Sesame on automotive AI toolchains, where I built graph optimization and quantization modules and deployed perception models onto embedded hardware for OEMs. I’ve spent a lot of time owning full pipelines from data processing and model training to serving, monitoring, and frontend integration. What makes me stand out is that I treat AI as an engineering problem, not a research exercise. I focus on reliability, low latency, and clean design, and I’m used to taking complex workflows and turning them into something fast, stable, and easy for users to depend on. I enjoy solving real problems with technology, and I’m comfortable taking responsibility for systems end to end.

Available to hire

Zhilin Xu said:

Here’s the same intro with no em dashes, written in a natural, client-ready voice:

My name is Zhilin Xu, and I’m a machine learning engineer who focuses on building AI systems that actually run in production. Over the last several years I’ve designed and shipped large-scale ML and LLM solutions for enterprise products. That includes real-time diagnostic intelligence at AppDynamics, multi-agent conversational systems for education, and high-performance model deployment using Python, PyTorch, FastAPI, and cloud tools.

Before that, I worked at Black Sesame on automotive AI toolchains, where I built graph optimization and quantization modules and deployed perception models onto embedded hardware for OEMs. I’ve spent a lot of time owning full pipelines from data processing and model training to serving, monitoring, and frontend integration.

What makes me stand out is that I treat AI as an engineering problem, not a research exercise. I focus on reliability, low latency, and clean design, and I’m used to taking complex workflows and turning them into something fast, stable, and easy for users to depend on. I enjoy solving real problems with technology, and I’m comfortable taking responsibility for systems end to end.

See more

Experience Level

Expert
Expert
Expert
Expert
Expert
Expert
Expert
Intermediate
Intermediate
See more

Language

English
Fluent

Work Experience

Machine Learning Engineer at AppDynamics
August 1, 2022 - November 27, 2025
Engineered an agentic AI system for real-time root-cause analysis by integrating anomaly detection, causal reasoning, and an LLM-based narrative synthesis to turn observability data into actionable insights. Built a RAG pipeline grounding LLM reasoning with enterprise telemetry using PostgreSQL + pgvector for contextual retrieval and fact-checked explanations across metrics, traces, and logs. Extended LangChain/LangGraph orchestration with MCP connectors to standardize tool interactions and support modular multi-agent workflows. Developed FastAPI-based microservices for anomaly detection, causal inference, and AppDynamics API integration, with OpenTelemetry, Redis, and Kafka/Flink for real-time ingestion and correlation. Created an interactive React + TypeScript dashboard for live RCA visualization with sub-second latency; deployments via Docker to AWS ECS/Lambda; CI/CD with GitHub Actions and CodePipeline.
AI Platform Engineer at Black Sesame Technologies Inc
August 1, 2022 - August 1, 2022
Implemented graph-optimization and INT8/FP16 quantization to boost inference throughput under edge power and thermal constraints. Designed a runtime scheduling layer in C++ under QNX RTOS and AUTOSAR to coordinate NPU–CPU workloads for multi-sensor fusion. Extended Shanhai AI Toolchain with ONNX/TensorFlow/PyTorch importers, implementing graph parsers, layer fusion, and tensor memory reuse for unified model conversion and optimized runtime execution across CPU, GPU, and NPU on QNX RTOS. Built benchmarking and validation pipelines with CUDA profiling and power measurement tools to enable OEM testing. Automated builds, tests, and deployments with GitHub/Jenkins CI/CD.
Software Engineer at Cornami, Inc.
April 1, 2021 - April 1, 2021
Designed and developed an AI framework to drive Cornami's low-power, high-performance AI chip for training and inference, accelerating ML models by implementing hardware-specific ML operators for large-scale batched data and streaming inputs. Automated ML operator interface generation during compilation and established test suites to validate against ONNX and JITlib benchmarks, reducing manual validation effort.
Software Engineer Intern at KLA
August 1, 2018 - August 1, 2018
Wafer information processing automation and acceleration; automated wafer defect matching, data statistics generation, and data pattern visualization to gain insight into Surfscan SPS and SP7 system characteristics.
Research Assistant at University of Waterloo
December 1, 2016 - December 1, 2016
Robot communication development: built communication infrastructure in C, C++, and Java to simultaneously control multiple robots; incorporated stream-based processing for real-time voice input into robot systems; developed cloud and Android applications to facilitate multi-channel access to robots.
Software Engineer Intern at IBM
December 1, 2015 - December 1, 2015
Code review application development; integrated Orion compare widget into IBM's Rational Team Concert's code review web application to enable in-RTC code reviews; resolved cross-stream code-change merge discrepancies; fixed 100+ code-review bugs and led UI/UX discussions across teams.
Software Engineer Intern at Zynga
April 1, 2015 - April 1, 2015
Android game development; updated game UI and contributed to Word Streak: Words With Friends; implemented a one-way friending algorithm to improve matchmaking and user socialization.
Software Engineer Intern at InfoMax Technologies Corporation
August 1, 2014 - August 1, 2014
Digital pen solution development; built web and Android applications to register handwritten forms using digital pens.
Database Administrator Intern at TELUS
December 1, 2013 - December 1, 2013
Database administration tasks supporting data management and operations.
Software Engineer Intern at TELUS
April 1, 2013 - April 1, 2013
Software development internship contributing to backend and tooling efforts.

Education

Master's degree at University of Southern California
August 1, 2017 - May 1, 2019
Bachelor's degree at University of Waterloo
September 1, 2012 - May 1, 2017

Qualifications

Add your qualifications or awards here.

Industry Experience

Software & Internet, Computers & Electronics, Media & Entertainment