Backend practice with small deployed APIs and honest limitations
I have not held a paid Backend Engineer title yet. What I can show is small-scale API work, local backend experiments, and documentation that clearly states what is missing.- Express API work: Car-Match deployed on Render with MongoDB Atlas and JWT auth.
- Backend experiments: FastAPI work for Convo-AI plus smaller schema and queue reps.
- Documentation: READMEs and issues list cold starts, missing features, and rough edges.
Clear scope, upfront
What I have
- Personal APIs deployed on small hosting providers for demonstration purposes.
- Local FastAPI experiments to learn backend service patterns.
- Documentation that lists missing features and risks.
What I am still working toward
- Ownership of production backend systems or high-traffic services.
- Tracing, alerting, and stronger observability beyond small-scale logs.
- Large-scale data modeling and migrations.
What I’m doing next
- Auth hardening (refresh tokens, auditing, device trust).
- Schema migration practice for relational databases.
- Async job queues and caching experiments.
Where I actually spend time
Personal deployments (public, small-scale)
This is the most concrete backend work I can point to right now: small public APIs with the tradeoffs clearly documented.
- Node and Express APIs for Car-Match, deployed on Render with MongoDB Atlas.
- Free-tier cold starts can take several minutes; this limitation is documented in the README.
- README documentation lists missing features such as authentication hardening, load testing, and observability.
Local experiments (learning)
I also use smaller local experiments to learn backend patterns without pretending they are more mature than they are.
- FastAPI experiments for Convo-AI to understand Python-based backend services.
- AI-assisted pair programming in development sessions; prompts and edits are included in the repository for traceability.
What I can show today
Car-Match API
Stack: Express, MongoDB Atlas, JWT, Render
Purpose: CRUD routes for matching users, forums, messaging, and auth flows.
- Limitations: free-tier cold starts, no rate limiting, limited structured logging, and no serious load testing yet.
- Docs: Postman collections and README tables instead of generated OpenAPI docs.
Proof links: Case study, GitHub
Secrets Management tutorial
Stack: demo frontend + documented backend/security patterns
Purpose: show secure versus insecure handling of secrets in a way beginners can inspect.
- Limitations: educational demo only, not a production backend service.
- Disclosure: large portions were scaffolded with ChatGPT and Copilot before manual annotation.
Proof links: EthicsFrontEndDemo repo, EthicsFrontEndDemo Live demo
Stacks I reach for
Each repository documents areas of strength and areas where AI tools or tutorials were heavily used.
Gaps I’m working through
- Designing auth flows that include refresh tokens, device trust, and auditing.
- Schema design + migrations for relational databases (beyond toy schemas).
- Observability and tracing for Express/FastAPI services.
- Performance profiling, caching strategies, and async job processing.
If you mentor junior backend engineers in these topics, I am open to discussing opportunities.