nky@devops:~$

Hello, world. I'm

Nkechi Anna Ahanonye

Cloud & DevOps Engineer / AWS · Ansible · Terraform · K8s

I turn 3 AM-breaking deployments into 1-min pipelines with AWS + Ansible + Terraform.
No more SSH-ing into prod. Now building security-first AI agents on Amazon Bedrock AgentCore.
AI Governance on AWS certified. Training 30+ African DevOps engineers.

View Pipeline Proof (2m 11s → 1m 18s) GitHub: ansible_practical
1m 18s Deploy Time (was manual)
40+ IaC Repos Audited
AgentCore Security-First AI Agents
.github/workflows/ansible-ci.yml 
name: Ansible CI/CD Pipeline
on: [push]
jobs:
  lint-and-dry-run:
    runs-on: ubuntu-latest
    steps:
      - uses: actions/checkout@v4
      - name: yamllint
        run: yamllint .
      - name: ansible-lint
        run: ansible-lint
      - name: syntax-check
        run: ansible-playbook --syntax-check site.yml
      - name: dry-run
        run: ansible-playbook --check --diff site.yml
  deploy:
    needs: lint-and-dry-run
    runs-on: ubuntu-latest
    steps:
      - run: ansible-playbook site.yml
# Stop SSH-ing into prod.
scroll
01 / about
Nkechi Ahanonye — Cloud & DevOps Engineer
📍 Lagos, Nigeria
Open to work
🚀 1-min Automated Pipelines
🛠 Infrastructure as Code (IaC)
☸ Kubernetes Orchestration
🤖 Security-First AI Agents

Engineering reliability. Automating growth.

My journey in tech started 15 years ago running a high-pressure cybercafe. That experience taught me a fundamental truth: real users don't care about your stack—they care that the system works. Downtime = lost revenue immediately.

I moved from physical hardware to cloud, obsessed with one thing: Reliability. Today I help teams stop SSH-ing into prod. I build 1-min automated pipelines: Lint → Dry-Run → Deploy with GitHub Actions + Ansible + Terraform.

In 2026 I took that same discipline to Agentic AI: security-first agents on Amazon Bedrock AgentCore — SigV4 and OAuth2 M2M tool authentication, per-request tool authorization via Amazon Verified Permissions, and agents architected to decline rather than guess. AI Governance on AWS certified (BeSA Agentic AI Batch 10).

My approach:

  • Efficiency: Terraform + Ansible slash setup from hours to minutes. Cut deployment time by 60%.
  • Stability: 6 checks before prod — yamllint, ansible-lint, shellcheck, syntax-check, --check --diff. No more 3 AM firefights.
  • Documentation: Mentor-grade runbooks that cut new member ramp-up by 30%.

Beyond engineering, I teach: a 15-module Ansible masterclass, a free Ansible Lab, and a 17-repo Blueprint series used by engineers across Africa. I co-mentor 30+ engineers at DevOps Micro Internship (DMI) and audited 40+ IaC repos for production-grade quality.

If your team needs an engineer who can architect, automate, and document production infrastructure that scales without the stress — I am open to Remote roles.

⚡
Lint → Dry-Run → Deploy Template reused as team standard
🐳
Containerization Docker · Kubernetes (k3s)
🎓
Teaching & Leadership DMI Mentor · 15-Module Masterclass · 30+ Engineers Trained
🛡
AI Governance on AWS Certified · BeSA Agentic AI Batch 10
Download CV
02 / skills

The stack I
deploy with.

⚙

Ansible

Playbooks, roles, inventory, Vault, Galaxy — 15-Module Lab

Advanced
🏗

Terraform

Multi-tier AWS, VPC, state, modules, 33 resources

Advanced
☁

AWS

EC2, S3, VPC, IAM, RDS, ALB, Route 53, Lambda

Advanced
🚀

GitHub Actions

Lint → Dry-Run → Deploy, 2m 11s → 1m 18s proof

Advanced
☸

Kubernetes

k3s, 3-node cluster, self-healing, ingress

Advanced
🐳

Docker

Dockerfiles, Compose, multi-tier

Advanced
🐧

Linux

Ubuntu, Nginx, MySQL, shell scripting

Advanced
🔁

CI/CD & Monitoring

Datadog, Prometheus, Grafana, ArgoCD

Intermediate
📝

Technical Docs

Architecture, SOPs, runbooks, DMI training

Expert
🧠

Agentic Workflows

Bedrock AgentCore, Strands, Claude, MCP, Lambda — deployed agents

Production builds + AI Governance certified
03 / case studies

Solutions I've
delivered.

🤖
GitHub Repo ↗
Featured Bedrock AgentCore Security-First Strands + Claude

agentcode — Security-First AI Agent on Amazon Bedrock AgentCore

Challenge: AI agents that call tools with no real security model.
Solution: Strands agent running Claude Opus 4.6 on AWS Bedrock AgentCore — every tool call secured with a different auth pattern (SigV4, OAuth2 M2M), per-request tool authorization via Amazon Verified Permissions, architected to decline rather than guess.
Result: Deployed and tested on live AWS, documented with 64 real console screenshots. Built from the AWS re:Invent SEC307 security workshop.

✅ Deployed on AWS ✅ Zero Secrets in Code ✅ 64 Console Screenshots
GitHub Actions 2m 11s failed to 1m 18s passed
GitHub Repo ↗ Live Actions ↗
GitHub Actions Ansible Lint → Dry-Run → Deploy

Ansible CI/CD Pipeline — 2m 11s ❌ → 1m 18s ✅

Challenge: Manual SSH deploys, no checks, 3 AM firefights.
Solution: Built GitHub Actions pipeline with 6 guards: yamllint, ansible-lint, shellcheck, --syntax-check, --check --diff, auto-deploy. Template now reused as team standard.
Result: Run #1 failed at 2m 11s catching errors before prod. Run #2 passed at 1m 18s. Zero SSH.

✅ Real Terminal Proof ✅ 1m 18s Automated ✅ No SSH Needed
☸
GitHub Repo ↗
Kubernetes AWS EKS Multi-Agent

Production-Ready AI Agents on EKS

Challenge: Multi-agent systems get complex and expensive fast.
Solution: 7 agents (customer, supervisor, RAG, memory, knowledge-graph, MCP, strands) deployed to both minikube and a real AWS EKS cluster, with cost guards from the first test.
Result: Entire project ran at $0.55 of a $1.00 budget — measured, not estimated.

✅ 7 Agents ✅ Real EKS Cluster ✅ $0.55 Total Cost
⚡
GitHub Repo ↗
Lambda Step Functions EventBridge

Serverless Agentic AI — Choreography vs Orchestration

Challenge: Choosing how agents coordinate on AWS.
Solution: Built both patterns — EventBridge choreography and Step Functions orchestration — with Lambda agents in Python.
Result: Two working architectures with honest trade-offs documented, not just slides.

✅ Serverless ✅ 2 Patterns Compared ✅ Real Deployments
🛒
GitHub Repo ↗ Live Store ↗
Kubernetes ArgoCD GitOps

CloudMart — GitOps E-Commerce on Kubernetes

Challenge: E-commerce deploys that drift without GitOps discipline.
Solution: Cloud-native store on Minikube — Docker, ArgoCD, NGINX Ingress, FastAPI, PostgreSQL 15, with a Jumia-style UI, real images, Naira prices and login/signup.
Result: Full store, GitOps-synced, live demo anyone can click.

✅ ArgoCD GitOps ✅ FastAPI + PostgreSQL ✅ Live Demo
🗳
GitHub Repo ↗
Hackaholics 7.0 Civic Tech Trust Platform

Ballot Trust 360 — Nigerian Election Trust Suite

Challenge: Every Nigerian's election pain is trust — accreditation failures, altered result sheets.
Solution: Complete civic trust suite: automatic accreditation, EC8A result-sheet verification, AI-generated fictional candidates.
Result: All-Nigerian Edition built for Hackaholics 7.0 — a trust demo, not a toy.

✅ Hackathon Build ✅ Civic Trust ✅ All Nigerian Edition
04 / how I deploy

My deployment
pipeline.

Stop SSH-ing into prod. Every project follows Lint → Dry-Run → Deploy. Same template that went 2m 11s ❌ → 1m 18s ✅

🔍

Lint

Catch errors before prod

yamllint . && ansible-lint && shellcheck *.sh
→
🧪

Dry-Run

See what would change, without changing

ansible-playbook --syntax-check --check --diff site.yml
→
🐙

Git Push → GitHub Actions

Automated guardrails on every push

git push origin main
→
🏗

Terraform Provision

VPC, SG, IAM, ALB — 33 resources

terraform apply -auto-approve
→
⚙

Ansible Deploy

Idempotent config, zero SSH

ansible-playbook site.yml
→
✅

Production Live — 1m 18s

No 3 AM calls. Runbook in Notion.

https://github.com/nkydigitech/ansible_practical/actions
05 / contact

Let's build
something reliable.

I help teams stop SSH-ing into prod. If you need 1-min pipelines that don't break at 3 AM, let's talk. Open to remote roles.

✉ nahanonye@gmail.com 🐙 github.com/nkydigitech — 99 repos 💼 linkedin.com/in/nkechiahanonye