Building systems that earn trust, and keeping the path from idea to production clear.
|
Portfolio |
|
Medium |
GitHub |
|
I like understanding the why before I build the what. I learn fast, ask thoughtful questions, adapt quickly when the path changes, and turn messy ideas into software that feels simple to use and solid to rely on.
Right now, that means building across production backend systems, cloud delivery, and product-facing workflows with Go, Python, PostgreSQL, AWS, Terraform, and GitHub Actions, while bringing strong project-building and backend engineering discipline from Java and Spring Boot.
| Current role | Focus | Writing |
|---|---|---|
| Software Engineer at Dexaminds | Patient-care platform, async workflows, backend reliability | Enterprise AI, runtimes, developer tooling |
- Building reliable services and async workflows with explicit contracts, idempotency, observability, and recovery paths.
- Shipping across backend, infrastructure, and product surfaces instead of treating them as separate silos.
- Using AI tools deliberately for exploration, scaffolding, and acceleration while keeping architecture, tradeoffs, and review under direct engineering control.
- Staying current with runtime shifts, AI product patterns, and developer tooling changes through hands-on work, technical writing, and continuous experimentation.
| Area | Proof |
|---|---|
| Patient-care platform | 52.3s ingestion, 48.1s CDIS, 0.2s handoff gap, 100.6s end-to-end runtime |
| Reliability patterns | 24-hour idempotency, KMS-backed versioned storage, tombstone-first deletes, DLQ-backed recovery |
| ResumeCanvas | 95.2% test coverage, 129 tests, 220 assertions, 60% query load reduction via 7 indexes |
| StageWay | 100% API test coverage, 12 Flyway migrations, 3 role-based permission levels |
| Internship impact | Fixed 15+ production bugs across frontend and backend and pushed verified fixes |
Codex · Claude · NotebookLM · Cursor
I use AI tooling to speed up drafting, test scaffolding, exploration, and synthesis. I keep system design, code review, architecture decisions, and release judgment human and explicit.
- I am highly curious about system design and architecture, especially where real delivery constraints force better decisions than theory alone.
- Hands-on production work sharpened how I reason about boundaries, failure modes, contracts, and long-term maintainability.
- I communicate clearly, ask thoughtful questions early, and adapt quickly when requirements, constraints, or rollout paths change.
- I enjoy understanding market and engineering shifts, then translating them into practical decisions about products, runtimes, and developer workflows.
Production work at Dexaminds spanning ingestion and CDIS, with ownership across service boundaries, workflow contracts, and AWS delivery paths.
- Built independently deployable Go services for control-plane and downstream processing.
- Shipped async workflows with
S3,EventBridge,Step Functions,SQS/DLQ,ECS,PostgreSQL, andTerraform. - Hardened the platform with idempotency, audit-safe storage, lineage-aware event history, and recovery paths for stale or stuck work.
AI-powered resume builder focused on performance, ATS relevance, and production-grade user experience.
- Built with
Next.js,TypeScript,MongoDB,Bun,NextAuth.js,Zod, andPuppeteer. - Added
8AI features,5resume templates, public sharing, and PDF export. - Reached
95.2%test coverage and cut query load by60%through targeted indexing. - Live: resumecanvas.live
Full-stack event management platform built with strong backend structure and delivery discipline.
- Built with
Spring Boot,PostgreSQL,Next.js,Docker,OAuth2,JWT,Flyway, andJPA/Hibernate. - Implemented
3permission roles,12migrations, QR check-ins, analytics, and waitlist workflows. - Maintained
100%API test coverage while keeping the architecture clean and deployment-friendly. - Live: stage-way.vercel.app
I write when I want to clarify a shift, not just consume it. Recent pieces published on Dexaminds Medium:
- Why enterprise AI is moving from chat interfaces to workspaces
- Why JavaScript runtimes are being rethought: from Node.js to Bun
Public GitHub signals from my main profile, kept here as supporting proof rather than the main story.
Open to software engineering roles where backend ownership, product thinking, cloud delivery, and engineering judgment all matter.


