What I can build, and the proof
Not a tech-stack list — the problems I can solve, each tagged by maturity and linked to evidence you can open: a project, a live feature, writing, or the repo.
Full-Stack & DevOps, end to end
The broad, proven base: I design, build, ship, and operate production systems across the frontend, backend, data, and cloud.
Software Architecture
I choose the architecture the problem needs — from a modular monolith to event-driven microservices — and keep the dependency direction honest.
- Modular Monolith
- Microservices
- Event-Driven Architecture
- Clean Architecture
- Domain-Driven Design
- CQRS
- Vertical Slice
Frontend / Web Engineering
This very site is the proof: Next.js App Router with Server Components, bilingual routing, motion, and a Lighthouse-first budget.
- Next.js App Router
- React Server Components
- SSR / ISR
- TypeScript
- Tailwind / Design Systems
- i18n (EN/VI)
- Core Web Vitals / SEO
- Accessibility
Database Engineering
Polyglot persistence done deliberately: Postgres for the relational core, Mongo for documents/events, Redis for nonce/rate-limit — plus full-text search on Postgres.
- PostgreSQL
- MongoDB
- Redis
- SQL Server
- Schema Design
- Indexing
- Query Optimization
- Full-Text Search (tsvector/GIN)
Cloud & Infrastructure
I run real workloads on AWS (ECS, S3, RDS, Aurora, CloudWatch) and ship this stack across Vercel, Render, Neon and Upstash — containerised and reverse-proxied.
- AWS (ECS/S3/RDS/Aurora/CloudWatch)
- Docker
- Nginx / Reverse Proxy
- Load Balancing
- CDN
- Object Storage
- Infrastructure as Code
DevOps & CI/CD
Every push runs the gates: lint, typecheck, test, build, then container build and deploy. Pipelines are documented, not improvised.
- GitHub Actions
- GitLab CI
- Azure DevOps
- Docker Build Pipelines
- Deployment Automation
- Monitoring & Alerting
- Log Aggregation
Production Readiness
Shipping is the start, not the finish. Health probes, structured logs, rate limiting, audit trails and a scaling story — built in, not bolted on.
- Monitoring
- Logging
- Tracing
- Backup & Disaster Recovery
- Rate Limiting
- High Availability
- Scalability
- Health Probes
Where I go deeper than most
Two differentiators with real proof: an encrypted-auth security stack running live on this site, and production BIM automation for the construction industry.
Security Engineering
The crown jewel runs live on this site: an encrypted request/response handshake (ECDH → HKDF → AES-256-GCM) with forward secrecy and replay protection, on top of Argon2id, rotating refresh tokens with reuse-detection, and TOTP MFA.
- Authentication
- Authorization (RBAC/ABAC)
- JWT
- OAuth2 / OpenID Connect
- OWASP Top 10
- Secret Management
- API Security
- Encryption (ECDH/HKDF/AES-GCM)
BIM & AEC Technology
My distinctive edge: production BIM automation across Revit, Etabs and Tekla — a tool ecosystem with 150+ tools and 30,000+ uses, plus an open-source Revit MCP SDK.
- Revit API
- APS (Autodesk Platform Services)
- BIM Automation
- Construction Technology
- Digital Twin
- BIM Data Pipelines
The parts most full-stack engineers skip
Data pipelines, system integration, automation, and the engineering discipline that keeps it all maintainable.
Data Engineering
I build the pipelines most full-stack engineers skip: AI-powered ETL over construction drawings, async crawlers, and document processing that survives 6,000-page inputs.
- ETL
- Data Pipelines
- Web Crawling (Crawl4AI)
- Data Warehouse / Lake
- Data Processing
- Data Validation
Enterprise Integration
Systems rarely live alone. I connect them with REST, webhooks, message queues and event buses — and a JSON-RPC contract between Revit and external apps.
- REST API
- GraphQL
- Webhooks
- Message Queue
- Event Bus
- Third-party Integration
Automation Engineering
Repetitive work is a bug. I automate engineering workflows, approval processes and desktop/CAD tasks — data-driven, not hard-coded.
- Workflow Automation
- Process Automation
- AI Workflow
- Browser Automation
- Desktop Automation
Engineering Practices
How I work shows in this repo: SOLID, design patterns, tested code, ADRs for every non-trivial decision, and conventional commits behind the gates.
- SOLID
- Design Patterns
- Unit Testing
- Integration Testing
- Code Review
- Technical Documentation (ADRs)
Learning in the open
An honest in-progress area — building real RAG and agent systems, not claiming mastery.
AI Engineering
An honest in-progress area. I'm building real RAG and agent systems — hybrid retrieval, MCP tooling, embeddings — in Knowlex and the Revit MCP SDK, and learning in the open.
- RAG
- AI Agents
- Multi-Agent Systems
- MCP (Model Context Protocol)
- Semantic Chunking
- Embeddings / Vector DB
- Prompt Engineering
- Tool Calling