Stephen Onyango Roy

DevOps & Platform Engineer | Nairobi, Kenya

Production Kubernetes, AWS, and Azure operations—with a focus on self-hosted LLM infrastructure.

Download ResumeContact Stephen

About Me

DevOps & Platform Engineer with 5+ years running production workloads on Kubernetes, AWS, and Azure across Kenya, Uganda, and Burkina Faso. I lead infrastructure automation, cloud governance, and reliable operations for business-critical workloads.

My current focus is self-hosted LLM infrastructure: an inference and RAG stack with Ollama, LangChain, LangGraph, and ChromaDB, packaged with Helm and monitored with Prometheus and Grafana. I have reduced cloud spend by more than 66% and led Terraform adoption across 1,000+ AWS resources.

66%+
AWS spend reduction
1,000+
AWS resources in Terraform

Current Focus

  • Self-hosted LLM inference & RAG
  • Production observability & incident response
  • Multi-Cloud Architecture Design

Skills Matrix

Cloud & Infrastructure
AWS (EC2, EKS, Lambda, RDS, S3, VPC, IAM, CloudWatch, SQS, SNS)
Azure
GCP
Multi-account & multi-cloud operations
Kubernetes & Containers
Amazon EKS
Kubernetes
Docker
ArgoCD & GitOps
Infrastructure as Code
Terraform
CloudFormation
Ansible
Automation & CI/CD
Python & Bash
GitHub Actions (OIDC)
AWS Lambda & EventBridge
API Gateway
Reliability & Observability
Prometheus & Grafana
Loki & Promtail
CloudWatch
Incident response & root-cause analysis
Security & Development
IAM, RBAC & VPC
SSO & secrets management
Vault
JavaScript & TypeScript
LLM & RAG
Ollama
LangChain & LangGraph
ChromaDB
Helm

Featured Projects

Self-Hosted LLM Serving & RAG Platform (Demo Prototype)
Demo
Demo prototype: a self-hosted LLM inference and RAG platform on Kubernetes, designed to run without cloud dependency on 8 GB RAM.

Technologies

Ollama
FastAPI
ChromaDB
LangChain
Helm
Prometheus
Grafana

Key Features

  • LLM inference
  • Retrieval-augmented generation
  • Helm packaging
  • Monitoring & CI pipeline
Explore live demo
Cloud Command Hub (Demo Prototype)
Demo
Demo prototype: a multi-account AWS observability dashboard for fleet health, security posture, cost attribution, and SLA tracking.

Technologies

AWS
Python
FastAPI
EKS
GitHub Actions
Prometheus
Loki

Key Features

  • EC2/EKS fleet health
  • IAM & GuardDuty visibility
  • FinOps cost attribution
  • Metrics & log aggregation
Explore live demo

Experience Timeline

DevOps Engineer

Living Goods

Apr 2022 – Present
Kenya, Uganda & Burkina Faso

Key Achievements

  • Reduced AWS expenditure from approximately $60K to below $20K per month—over 66% savings, or more than $480K annualised.
  • Led Terraform adoption, importing 1,000+ AWS resources across development, staging, production, and data warehouse environments.
  • Led migration from self-managed Kubernetes (kOps) to Amazon EKS and from a single AWS account to a governed multi-account AWS Organization.
  • Implemented SSO-based AWS access, eliminating reliance on long-lived access keys.

Core Responsibilities

  • Operate production AWS, Azure, Kubernetes, Linux, and Windows workloads across three countries.
  • Manage AWS services including EC2, Lambda, RDS, S3, IAM, VPC, CloudWatch, SQS, and SNS.
  • Manage Kubernetes state across environments with ArgoCD and GitOps.
  • Built observability with Prometheus, Loki/Promtail, Uptime Kuma, and AWS Cost and Usage Reports.
  • Automate operations with Terraform, Lambda, EventBridge, API Gateway, Python, Bash, and GitHub Actions OIDC.
  • Resolve production incidents across applications, APIs, Kubernetes, infrastructure, networking, and databases.

DevOps Engineer

KodeKloud — Project Nautilus

Feb 2020 – Mar 2022

Key Achievements

  • Administered multi-server Linux environments and built Docker and Kubernetes workloads.
  • Troubleshot networking, storage, and runtime issues.
  • Automated infrastructure with Ansible and built Jenkins CI/CD pipelines.

Customer Technical Support Engineer

Raj Ushanga House Limited

Dec 2018 – Mar 2022

Key Achievements

  • Provided L2/L3 support for AWS- and Azure-hosted IoT platforms across Linux, cloud infrastructure, networking, and APIs.
  • Investigated production availability, performance, and distributed-system issues.
  • Created runbooks and troubleshooting procedures adopted across the support team.

Software Engineer

Andela

Feb 2019 – Mar 2020

Key Achievements

  • Developed, tested, debugged, and maintained software applications against functional and technical requirements.
  • Investigated defects, implemented fixes, and collaborated across the software development lifecycle.

Certifications & Education

Certifications

AWS Certified DevOps Engineer – Professional

Amazon Web Services

Certified Kubernetes Administrator (CKA)

CNCF

AWS Certified Developer – Associate

Amazon Web Services

AWS Certified Cloud Practitioner

Amazon Web Services

Kubernetes and Cloud Native Associate (KCNA)

CNCF

ITIL Foundations

ITIL

Education

BSc, Information Technology

Jomo Kenyatta University of Agriculture and Technology (JKUAT)

Professional Profiles

Find me on these platforms.

LinkedInGitHubMedium

Get In Touch

Based in Nairobi, Kenya · +254 706 519 258. For opportunities and technical conversations, reach me at steve.onyangoroy@gmail.com.