Bridging software engineering and cloud infrastructure to build reliable, scalable systems.
Computer Science graduate with a backend, data and AI foundation and a cloud and DevOps focus. I take a project from design to production and keep it running: architecture first, small tested increments, security at every layer, automated delivery, observability, and documentation that makes the work easy to review and hand over.
My cloud graduation project first, then what I am building now to learn observability, reliability and AWS.
NTI Cloud Operations · Graduation project
Deploy a Multi-Tier Web Application
A multi-tier web application infrastructure designed around scalability, reliability, security, and repeatable deployments. Application and data components stay private behind a controlled public entry point.
Engineering focus: make the architecture understandable, keep access limited, and verify each layer before integrating the next.
Architecture summary for portfolio purposes. Refer to the repository for implementation details.
In progress
PulseTrace: microservices with observability foundations
A local-first microservices platform that simulates an e-commerce order flow across three independent Node.js services (API Gateway, Order, and Payment). Each service owns its own PostgreSQL database, and the stack runs under Docker Compose with health checks. Only the gateway is exposed.
Engineering focus: structured logs with request IDs, Prometheus metrics from every service, a provisioned Grafana dashboard, and automated tests that cover the observability layer itself, not just the happy path.
Show the architecture and trace a request
Solid = implemented · Dashed = planned.
1 · Client
HTTP clientAttaches a request ID on every call
↓
2 · Entry
API GatewayExpress :3000 · sole public entry · routes & correlates
Client → Gateway (only public surface) → Order and/or Payment → each service's own database. Every hop carries the same request ID and emits structured logs (Pino) + Prometheus metrics, so one call can be reconstructed across three services.
Correlation
Structured JSON logs (Pino) carry a request ID across Gateway → Order → Payment.
Metrics
Request rate, latency histograms, error rate and business counters, scraped into a provisioned Grafana dashboard.
Testing
Unit, API, end-to-end, Docker/PostgreSQL integration, logging correlation and metrics tests.
Early stage: Module 1 (foundations) is in progress; monitoring, incident, automation and GitOps modules come next. Dashed tools are covered by the course and not yet applied here.
In progress · Coursera / Packt
AWS Cloud Projects
A living collection of hands-on AWS projects I rebuild myself while studying the AWS Cloud Projects course by Packt on Coursera. Each scenario becomes an implementation I design, provision, test and document: why this service, what fails, what it costs, how it could improve.
How each project is run
Per project
Capture requirements, design the architecture, implement with IaC (Terraform / CloudFormation), deploy, test, review security and cost, then keep the docs in sync.
Focus
Service interaction, provisioning, deployment and testing, security, scalability, maintainability and cost.
Organization
Each project has its own directory with README, docs, IaC and application code, with clear commits.
AWSTerraformCloudFormationIAMVPCComputeStorageSecurityMonitoringCI/CDServerlessAI / ML
Actively developed; projects are added one at a time as I progress through the course.
Completed · Software Testing course
Test design for a sailing-rules method
Tests for a Java method that decides whether sailing is available: the ship sails on weekends and every day in July, wind force must be 0–12, and sailing is only available below force 5. I applied four techniques and measured what each one covered.
Equivalence partitioning
Input and output partitions, including invalid wind values that raise exceptions, turned into 7 test cases.
Statement coverage
Measured with JaCoCo (about 87.5%, one missed branch), located the line on a control flow graph, and added a test to reach 100%.
Combinational testing
Cause-and-effect analysis: an 8-rule truth table reduced to 5 rules using don't-care conditions, with two new tests.
Random testing
Pick the expected result first, then generate inputs that force it, over 1,000,000 iterations, with and without a fault in the method.
Metrics, logs and traces: seeing how a system behaves and troubleshooting with evidence, applied step by step in PulseTrace.
Security & DevSecOps
Building security into the pipeline and the infrastructure, not adding it at the end.
MLOps & applied AI
How ML workflows connect with infrastructure, automation and observability.
Concepts first, tools second. A new tool still means learning its own syntax and limits, but the problem underneath is often one I already understand: declare desired state, isolate and package software, automate build and delivery, apply least privilege.
These notes are personal working documents written while learning, not polished documentation. Expect shorthand and unfinished sections.
$ ./contact.sh --send
Let's build something useful.
Open to graduate and junior opportunities in cloud, infrastructure, software engineering and technical support. I'm eager to learn, contribute and grow through practical work.