Terraform for AI Agents (8): End-to-End — research-agent-stack in One Apply
Stitching the seven modules into one repo, running terraform apply once, and watching a complete agent runtime — VPC, ECS, RDS, OpenSearch, …
9 min read · 1823 wordsAn editorial blog by Chen Kai
Long-form notes on machines, math, and the cloud.
A long-form notebook on machine learning, mathematics, and the cloud infrastructure that runs them.
Featured · Terraform Agents
Terraform for AI Agents (8): End-to-End — research-agent-stack in One Apply
Stitching the seven modules into one repo, running terraform apply once, and watching a complete agent runtime — VPC, ECS, RDS, OpenSearch, OSS, LLM gateway, SLS observability, cost alarms — come up in seven minutes. Real apply output, the …
Read the full piece →Ongoing series
Each one is a single argument unfolded chapter by chapter.
Aliyun Bailian
5
Aliyun Bailian
Bailian model platform: prompt engineering, fine-tuning, agents, and evaluation.
Aliyun PAI
5
Aliyun PAI
Production-grade ML on Alibaba Cloud — DSW, DLC, EAS, Designer, QuickStart, end-to-end.
Cloud Computing
8
Cloud Computing
Infrastructure, networking, and the platforms ML actually runs on.
Computer Fundamentals
6
Computer Fundamentals
OS, networking, compilers — the substrate beneath everything.
LeetCode Patterns
10
LeetCode Patterns
Algorithms by pattern, with worked solutions.
Linear Algebra
18
Linear Algebra
The geometry and computation that underlies all of ML.
Latest writing
Recent essays and deep dives, freshest first.
Terraform for AI Agents (7): Observability, SLS Dashboards, and Cost Alarms
Logs to SLS, traces to ARMS, metrics to CloudMonitor — all provisioned in HCL so a new env comes pre-instrumented. The four alarms that …
10 min read · 2069 words
Terraform for AI Agents (6): LLM Gateway and Secrets Management
Centralise LLM API access through one gateway: per-agent quotas, request logging, and zero secrets outside KMS. Terraform-provisioned API …
10 min read · 1970 words
Terraform for AI Agents (5): Storage — Vector, Relational, and Object Memory
An agent has three kinds of memory and they map onto three Aliyun services: PolarDB/RDS for sessions, OpenSearch (vector edition) or …
9 min read · 1792 words
Terraform for AI Agents (4): Compute — ECS, ACK, or Function Compute?
The three places an agent's main loop can live on Aliyun: a long-running ECS instance with pm2, a Kubernetes pod on ACK, or a Function …
9 min read · 1846 words
Terraform for AI Agents (3): A Reusable VPC and Security Baseline
The first reusable module — a three-zone VPC with public/private subnets, NAT egress, security groups layered by tier, and KMS keys per data …
10 min read · 1940 words
Terraform for AI Agents (2): Provider, Auth, and Remote State on OSS
Pinning the alicloud provider, picking between AK/SK, AssumeRole, and ECS RAM role auth, putting tfstate on OSS with Tablestore locking, and …
10 min read · 1947 words
Terraform for AI Agents (1): Why IaC Is the Only Sane Way to Ship Agents
Agent systems are a moving target — new tools, new memory stores, new regions every month. Manual console clicks don't survive the second …
8 min read · 1667 words
Aliyun PAI (5): Designer vs Model Gallery — When the GUIs Actually Earn Their Keep
PAI-Designer for tabular ML pipelines, Model Gallery for one-click open-source model deploy/fine-tune. The honest decision matrix for when …
5 min read · 1037 words