Python & AI software developer · Learning AWS architecture in public · Local-first engineering
I build software, experiment with cloud infrastructure, and document the decisions behind my work. This profile is a living engineering journal: it shows what I am learning, what I actually tested, where things failed, and how my approach improves over time.
Current focus (October–November 2026): a 30-day, hands-on journey toward the AWS Certified Solutions Architect – Associate (SAA-C03), with four AWS design domains and a growing evidence-based portfolio.
Exam scenario → Independent reasoning → Six-step conceptual analysis → Lab hypothesis → Local code + AI collaborator → Hand-drawn / draw.io architecture → Controlled AWS test → Evidence + reflection → Re-solve the question
I use AI to explain, prototype, test, and review—not to replace my architecture decisions. When a lab is too expensive or inaccessible, I label it design-only or local simulation rather than claim an AWS deployment.
| Domain | Exam weight | What I practice |
|---|---|---|
| D1: Secure Architectures | 30% | IAM, least privilege, encryption, trust boundaries |
| D2: Resilient Architectures | 26% | recovery, fault tolerance, decoupling, RTO/RPO |
| D3: High-Performing Architectures | 24% | storage, compute, data access, performance trade-offs |
| D4: Cost-Optimized Architectures | 20% | cost drivers, budgets, pricing trade-offs |
| Stage | Dates (2026) | Outcome | Status |
|---|---|---|---|
| Setup | Oct 16–18 | Accounts, safety, baseline assessment | 🟡 Planned |
| L1: Foundation | Oct 19–24 | IAM, S3, observation, budgets | 🟡 Planned |
| L2: Builder | Oct 26–31 | Python + Lambda + DynamoDB prototype | 🟡 Planned |
| L3: Architect | Nov 2–7 | SQS, DLQ, failure testing, trade-offs | 🟡 Planned |
| L4: Professional | Nov 9–14 | reviews, architecture alternatives, cost estimates | 🟡 Planned |
| Reflection | Nov 15 | practice assessment + portfolio review | 🟡 Planned |
No labs are marked complete yet. Exercise statuses live in the progress tracker; local tests and live AWS runs are never conflated.
- Q001 · Bicycle Location API — English / فارسی — an IAM-protected API Gateway → Lambda console/Terraform teaching kit, with manual execution, Draw.io worksheet and verified-cleanup checklist. Kit available; no live AWS run claimed.
- 30-day roadmap — four levels, 16 scoped exercises.
- Labs — one activity per domain per week.
- Question-to-Lab method — turn exam scenarios into engineering evidence.
- Architecture diagrams — my own editable draw.io sources and exports.
- Weekly reviews — decisions, mistakes, costs, and next steps.
- Cloud cost & security rules — local-first and controlled deployments.
- Getting started — reproduce the project and run validation.
Python · Git · Docker · FastAPI · Boto3 · AWS · Infrastructure as Code (learning) · draw.io · CI/CD · AI-assisted development
- Progress, not perfection: preserve early designs and later revisions.
- Reasoning, not memorization: explain why each architectural choice fits.
- Evidence, not screenshots alone: report what was actually tested.
- Cloud transparency: distinguish estimated cost, local behavior and AWS measurements.
- Human ownership: I draw and review architectures; AI tools support rather than replace my decisions.
Independent learning portfolio. Not an AWS-endorsed product or a claim of certification. All progress is recorded only after verification.