DEVOPS / MLOPS / ORACLE DATABASES
I build and operate cloud infrastructure, ML services, and Oracle databases, with a focus on reliability, automation, and observability. My research explores low-resource Ghanaian languages, computer vision, and accessible AI.
MSc Computer Science | DevOps & MLOps Engineer | Oracle Database Administrator
| Area | What I work on |
|---|---|
| Cloud & DevOps | Azure/AWS infrastructure, Kubernetes deployments, CI/CD, and workflow automation |
| MLOps | Reproducible training, evaluation, model packaging, serving, and monitoring |
| Oracle & DataOps | RMAN recovery, Data Guard, SQL tuning, migrations, and automated maintenance |
| Observability | Metrics, logs, dashboards, and alerting with Grafana, Prometheus, Elasticsearch, and Datadog |
- Ghanaian language technology: ASR and language detection for low-resource and accessibility use cases.
- Computer vision: CNN image classification in Akan Twi, including dataset curation and supervised training.
- ML infrastructure: Scalable cloud environments and CI/CD for deploying and operating ML services.
Other research and project interests include SMS spam detection in Swahili, speech recognition for people with speech impairments in Twi, and brain tumor classification using federated learning.
Build CI/CD workflows and containerized deployments on Azure/AWS, with an emphasis on security, observability, and reliable application and ML operations.
- Manage backups, recovery, and automated maintenance using RMAN.
- Tune SQL, indexes, and partitions with AWR/ASH analysis.
- Implement high availability and disaster recovery with Data Guard; coordinate patching and upgrades.
- Automate DBA tasks with PL/SQL, SQL*Plus, Bash, and PowerShell.
- Monitor databases with OEM/Cloud Control, Grafana, and Datadog.
- Investigated defects and supported internal and external teams.
- Performed regression testing, SQL data validation, and performance troubleshooting.
- Built Grafana dashboards and used Datadog to investigate bottlenecks.
- Automated repetitive support tasks with Bash.
| Discipline | Tools & technologies |
|---|---|
| Cloud & infrastructure | Azure, AWS, Docker, Kubernetes, Terraform, Ansible, Vagrant, Linux |
| Delivery & automation | GitHub Actions, Azure DevOps, Jenkins, Bash, PowerShell |
| ML & data | Python, PyTorch, TensorFlow, scikit-learn, Pandas, Polars, PySpark |
| ML lifecycle | MLflow, DVC, Prefect, Airflow, Kubeflow, KServe, Triton, FastAPI |
| Databases & validation | Oracle, SQL, PL/SQL, SQL*Plus, RMAN, Data Guard, Liquibase, Flyway, Great Expectations |
| Monitoring | Grafana, Prometheus, Elasticsearch, Datadog, OEM/Cloud Control |
A reference workflow for taking a model from source data to a monitored service:
flowchart LR
A[Data Sources] --> B[Ingestion & Validation]
B --> C[Feature Engineering]
C --> D[Train & Evaluate]
D --> E[Model Registry]
E --> F[CI/CD: Build / Test / Scan]
F --> G[Containerize & Deploy]
G --> H[Online Serving: REST and gRPC]
H --> I[Monitoring & Logging]
I --> J[Drift Detection & Alerts]
J --> D
Versioned data, reproducible pipelines, model comparison, and repeatable releases support each stage. Production services need testing, security checks, and observability throughout their lifecycle.
Advanced cloud infrastructure, ML pipeline orchestration, and privacy-preserving federated learning for healthcare applications.
Let's connect about cloud platforms, production ML, Oracle reliability, or language technology for Ghanaian communities.