Goh Chun Lin

Polyglot Systems Architect.

I build high-performance simulation engines and deterministic safety infrastructure for AI Agents.

Currently architecting SilverAi (Deterministic Safety for IoT & Agentic Systems). I am also a proud member of the .NET Foundation, bridging the gap between Enterprise Architecture and Modern AI.

Systems Architect

Chun Lin

I am Chun Lin, a Polyglot Systems Architect based in Singapore. My focus has shifted from building traditional web apps to architecting Agentic AI Infrastructure and IoT Safety Layers. I specialise in solving the "Last Mile" problem of deploying probabilistic LLMs into deterministic hardware environments.

With a background as a Senior Cloud Engineer and .NET Foundation member, I bring rigorous engineering standards to the chaotic world of AI. I do not just write code. I build systems that are resilient, observable, and safe.

Core Tech Stack:

  • Languages: Python, Go, C#
  • Simulation: Discrete Event Simulation (SNA)
  • Cloud: AWS, Azure
  • LLM Ops: LangChain, Semantic Kernel
  • Containers: Docker, Kubernetes
  • Databases: Aurora, DynamoDB, Pinecone

Featured Architecture

I build open-source infrastructure to solve critical engineering gaps in AI Safety and Cloud Performance.

SilverAi
Type: AI Safety Middleware
Stack: Python, Pydantic

A deterministic state-guardrail system for Agentic Hardware (IoT/Robotics). It intercepts and blocks hallucinated LLM commands that violate physical safety constraints (e.g. Battery limits, Thermal thresholds) before they reach the device API.
View on Github
SNA Simulation
Type: Discrete Event Sim
Stack: C#, .NET

A high-performance simulation engine used to model cost and latency in complex distributed systems. I use this to forecast the ROI and Operational Load of AI Agent deployments at scale.
View on Github
go-onedrive
Type: API Client SDK
Stack: Go (Golang)

A production-grade OneDrive SDK for Go. Gained worldwide adoption and community contributions. Demonstrates strict adherence to RESTful patterns and strongly-typed API design in Golang.
View on Github

Engineering Notes

Thoughts on Architecture, Safety, and the intersection of Physics & AI.

LLM First Principles

Stop treating LLMs as magic black boxes. To build robust systems, we must understand the 'physics' under the hood. I break down Parameter Counts, Context Windows, and Throughput using cloud engineering analogies (Stateless REST APIs, Database Engines) to help architects choose the right model for the job.
Read on Medium
A Kubernetes Lab for Massively Parallel .NET Parameter Sweeps

How I architected a distributed lab to run multiple concurrent Monte Carlo simulations. Using Kubernetes Jobs and .NET containers to solve the Embarrassingly Parallel problem of parameter sweeps without rewriting the core engine.
Read on Wordpress
Observing Orchard Core: Traces with Grafana Tempo and ADOT

Implementing the Third Pillar of observability (Distributed Tracing) using Grafana Tempo, ADOT, and OpenTelemetry. A deep dive into instrumenting .NET applications on AWS to visualize request flows across microservices.
Read on Wordpress

Career History

As a Senior Cloud Software Engineer, I lead the implementation of the next-generation AI Customer Assistant. I focus on reducing DDU (Cost of Product) by automating diagnostics while ensuring strict safety compliance for hardware control.

/AI Safety Architecture

Architected the Safety Middleware layer to prevent LLM hallucinations from triggering unsafe hardware states (Thermal/Battery protection).

/Agentic Orchestration

Implemented complex agent workflows using LangChain and Semantic Kernel, integrating with multiple LLMs (OpenAI and DeepSeek) for global markets.

/Cost & Latency Simulation

Utilised Discrete Event Simulation to forecast API costs and latency at scale, driving architectural decisions to reduce cloud spend.

/LLMOps and AI Engineering

Built LLMOps infrastructure on Amazon ECS, ensuring PII compliance and high availability for global rollouts.

Community & Speaking

I believe in sharing knowledge to elevate the engineering ecosystem. As a .NET Foundation member and frequent speaker, I bridge the gap between Enterprise Architecture and the new wave of AI Development.

Selected Talks & Workshops:

  1. K8s 上的模擬實驗室：自動化 .NET 參數掃描 (Hello World Dev Conference, Taipei)
  2. Observability in Orchard Core using CloudWatch: a gateway to Grafana (GrafanaCON Local Singapore)
  3. Integrating AWS CloudFormation with .NET Aspire (AWS User Group Singapore)
  4. Modernising Legacy Codebases with NDepend (.NET Conf)
  5. From Pkl to JSON: Infrastructure as Code Patterns (AWS User Group Singapore)
  6. Reality of the Clouds - Migrating Enterprise Systems to Azure (Azure Community Singapore)
  7. Unit Testing Data Layers in CI/CD Pipelines (Singapore .NET Developers)

Previous organiser of the Singapore .NET Developers Community (2017-2021).

Visit dotnet.sg

Creative Outlet

Engineering requires logic while drawing requires emotion. I believe maintaining a creative hobby keeps the mind flexible for architectural innovation.

My public drawing submissions are as follows.

  1. "The Promise of a Worldly Encounter" - Hoyoverse Fan Art
  2. SGOCF 2022 Art Exhibition - Min
  3. SGOCF 2023 Art Exhibition - Gacha

My drawing in SGOCF 2023.

Let's Build the Future

Open to consulting on AI Safety Architecture and Cloud Systems.
Message me on LinkedIn.

Copyright © 2025 Goh Chun Lin. All Rights Reserved.