DevSecOps · MLOps · AI Engineering Houston, TX · 10+ Years

Ugonna
B. Ekoro

Lead DevSecOps & MLOps Engineer

I architect and deliver secure, automated ML systems on AWS — from infrastructure provisioning with Terraform to production inference APIs, drift monitoring, RAG knowledge bases, and AI agents that operate infrastructure autonomously.

View Projects Experience LinkedIn ↗
10+ Years in DevSecOps
7 Certifications
3 Cloud Platforms
Ugonna B. Ekoro
01 — Expertise

What I Do

My work sits at the intersection of secure cloud infrastructure, machine learning deployment, and AI engineering. I've delivered across AWS, Azure, and OCI — in FedRAMP/GovCloud environments serving DoD and U.S. Treasury workloads.

⚙️

DevSecOps & CI/CD

Building and operating secure delivery pipelines with Jenkins, GitLab CI, Bamboo, and Azure DevOps. Integrating security tools (Fortify, SonarQube, Prisma Cloud, Nessus) across the full SDLC for continuous ATO compliance and NIST 800-53 alignment.

🤖

MLOps & AI Engineering

End-to-end ML platforms on AWS: training pipelines, experiment tracking with MLflow, real-time inference APIs, SageMaker Serverless deployment, drift detection with Evidently AI, RAG systems with OpenSearch, and AI agents via Bedrock.

🏗️

Cloud Infrastructure

Infrastructure as Code with Terraform across AWS, Azure, and OCI. Kubernetes cluster operations on Rancher, EKS, and AKS. Configuration management with Ansible. 15+ reusable Terraform modules reducing provisioning time by 40%.

🔐

Security & Compliance

FedRAMP High and GovCloud delivery for DoD and federal agency workloads. RHEL Satellite lifecycle management, STIG compliance, and security hardening. Public Trust clearance holder with hands-on ATO process experience.

📊

Monitoring & Observability

ML observability stacks monitoring API health, model accuracy, and data drift via custom CloudWatch metrics and SNS alerting — reducing MTTD for model degradation by 60%. Prometheus, Grafana, ELK, Fluentd, and Splunk for infrastructure observability.

🧠

AI Agents & RAG

Conversational AI agents using Bedrock Claude with tool calling (ReAct pattern). RAG knowledge base systems with hybrid vector and keyword search over OpenSearch. Sentence-transformers for semantic embeddings. Full pipeline from ingestion to generation.

02 — Technical Stack

Tools & Technologies

Cloud & AWS Services
EC2S3 IAMLambda BedrockSageMaker CloudWatchSNS DynamoDBAPI Gateway Route 53Azure OCIGovCloud
MLOps & AI
MLflowKubeflow KServeDVC Evidently AIOpenSearch RAGAI Agents Bedrock ClaudeLangChain
Machine Learning
PyTorchscikit-learn HuggingFaceBERT sentence-transformers pandasNumPy FlaskRandomForest
DevSecOps & CI/CD
JenkinsGitLab CI BambooAzure DevOps TerraformAnsible DockerKubernetes HelmRancher EKSAKS
Security & Compliance
FedRAMPNIST 800-53 FortifySonarQube Prisma CloudNessus Trend MicroSplunk STIGATO
Languages & Monitoring
PythonBash GroovyHCL YAMLSQL PrometheusGrafana ELK StackFluentd
03 — Experience

Career Timeline

Lead DevSecOps & MLOps Engineer
Oct 2023 – Present
T-Rex Solutions · FedRAMP/GovCloud · U.S. Federal Agency
  • Architected an end-to-end MLOps platform on AWS — cutting model deployment from multi-day manual processes to under 2 hours via Jenkins/SageMaker pipelines, Terraform infrastructure, and MLflow experiment tracking.
  • Built a production Flask REST API serving real-time ML inference, model explainability, and conversational AI endpoints integrated with AWS Bedrock (Claude 3 Haiku), secured via IAM role-based auth in a FedRAMP-compliant environment.
  • Built an ML observability stack monitoring API health, model accuracy, and data drift via custom CloudWatch metrics and SNS alerting — reducing MTTD for model degradation by ~60%.
  • Provisioned and standardized cloud infrastructure using 15+ reusable Terraform modules across AWS, Azure, and OCI — reducing environment provisioning time by ~40% and ensuring NIST 800-53 compliance.
  • Administered Kubernetes platforms (Rancher, EKS, AKS) with autoscaling, RBAC, and resource quotas supporting 99.9%+ platform uptime SLAs.
  • Led DevSecOps delivery for 5+ application workloads, championing IaC adoption and eliminating manual provisioning across 3 teams.
  • Integrated Fortify, SonarQube, Prisma Cloud, Nessus, and Trend Micro into CI/CD pipelines for continuous ATO compliance on Treasury workloads.
Senior DevSecOps Engineer
Apr 2022 – Sep 2023
L3Harris Technologies · DoD Hybrid Environment
  • Supported a DoD hybrid environment (on-prem + cloud), delivering secure and repeatable releases across classified and unclassified networks.
  • Built and operated CI/CD pipelines with Jenkins, Bamboo, and GitLab CI; led containerization with Kubernetes, Docker, Helm, JFrog Artifactory, and Nexus.
  • Administered NGINX and Apache HTTPD reverse proxies and managed CentOS VMs on VMware ESXi in on-premises environments.
  • Automated infrastructure deployments with Terraform and Ansible — reducing environment build time by 35% across dev/staging/prod.
  • Implemented monitoring and logging with Prometheus, Grafana, ELK/EFK, Fluentd, and Splunk — improving MTTD for production incidents by 45%.
Senior DevOps Engineer
Sep 2021 – Mar 2022
Oceus Networks · DoD Azure Government (FedRAMP High)
  • Delivered DevOps capabilities in a DoD Azure Government (FedRAMP High) environment using Jenkins, GitLab CI, Terraform, Ansible, and Kubernetes.
  • Implemented disaster recovery and performance monitoring for mission-critical DoD systems; established GitLab branching and release governance strategies.
  • Standardized Kubernetes deployments with Helm charts — enabling versioned releases, environment-specific configs, and reliable rollbacks.
DevOps Engineer
Feb 2017 – Aug 2021
Avienn Systems · AWS / Azure
  • Implemented CI/CD pipelines for microservices using Maven, Jenkins, SonarQube, Fortify, Kubernetes, Prometheus, and Grafana.
  • Provisioned AWS infrastructure (VPC, EC2, ECS, IAM, S3, CodeDeploy, CloudFront) using Terraform and CloudFormation.
  • Automated platform provisioning across AWS and Azure with Terraform and Ansible; developed Bash and Groovy automation scripts for build, deployment, and operational workflows.
  • Administered RHEL systems and automated deployments with Ansible, supporting repeatable and auditable production environments.
04 — Projects

Featured Work

A complete MLOps platform built from scratch on AWS — three machine learning projects unified under a single Flask API, automated by seven Jenkins pipelines, monitored by Evidently AI and CloudWatch, and operated by an AI agent accessible from Telegram.

Personal Project · AWS · 2024–2026
MLOps Platform on AWS
End-to-end machine learning platform demonstrating the full ML engineering lifecycle: infrastructure as code, CI/CD automation, model training, drift monitoring, RAG knowledge base, and an autonomous AI operations agent.
01 / Traditional ML
Titanic Inference API
RandomForestClassifier trained on Titanic passenger data with automated drift simulation, quality gates (0.75 accuracy threshold), MLflow experiment tracking, and a Flask REST API serving real-time predictions. 7 Jenkins CI/CD pipelines automating the full lifecycle.
scikit-learnFlask MLflowJenkins TerraformCloudWatch
02 / Deep Learning
BERT Support Ticket Classifier
Fine-tuned bert-base-uncased on 8,867 banking support tickets achieving 92% test accuracy. Gradient accumulation resolved memory constraints during CPU training. Deployed to SageMaker Serverless Inference for zero-idle-cost production serving.
PyTorchHuggingFace SageMakerboto3
03 / RAG System
Knowledge Base + Hybrid Search
Retrieval Augmented Generation system indexing DevOps and MLOps documentation in OpenSearch. Hybrid search combining dense kNN vectors with BM25 keyword search via Reciprocal Rank Fusion. Claude generates grounded answers with source citations.
OpenSearchsentence-transformers Bedrock ClaudeRRF
04 / AI Agent
OpsBot — Operations Agent on Lambda
Conversational AI agent (ReAct pattern) on AWS Lambda with 14 tools: EC2 power control, CloudWatch monitoring, GitHub activity tracking, service health checks, and RAG knowledge base search. Controlled via Telegram 24/7.
AWS LambdaBedrock Haiku DynamoDBTelegram API
92% BERT Accuracy
<2hrs Deploy Cycle
14 Agent Tools
7 CI/CD Pipelines
↗ GitHub · doubled192
05 — Credentials

Certifications

☁️
AWS Certified DevOps Engineer – Professional
Amazon Web Services
🤖
AWS Certified Machine Learning Engineer – Associate
Amazon Web Services
💻
AWS Certified Developer – Associate
Amazon Web Services
🏗️
HashiCorp Certified: Terraform Associate (HCTA-003)
HashiCorp
🔐
CompTIA Security+
CompTIA
🐧
CompTIA Linux+
CompTIA
🖥️
CompTIA A+
CompTIA
06 — Contact

Let's Connect

Feel free to reach out via LinkedIn or email — whether it's about MLOps architecture, AI agent systems, FedRAMP delivery, or anything in between.

✉️  Ugonna.ekoro@gmail.com 💼  LinkedIn ⌨️  GitHub · doubled192