Code that ships. Engineering that scales.
Production-ready AI platforms — React frontends, FastAPI backends, Docker, Kubernetes, and AWS.
AI + Backend systems · AWS CCP · Remote-ready
Clean architecture, shipped systems, and measurable outcomes — from AI search to real-time collaboration.
AI-Powered Career Operating System. Analyzes resumes against job descriptions using semantic matching and generates personalized improvement recommendations.
Real-Time AI Chat Platform with interruption-safe streaming, conversation persistence, and responsive chat UI.
Collaborative AI workspace for documents — semantic search, real-time editing, and workspace-level access control.
Two production failures, root causes, and the fixes that stuck.
Expired sessions locked the dashboard in an infinite redirect spinner.
Middleware trusted cookie presence while the API returned 401 — login and dashboard bounced forever.
Clear the invalid token cookie on JWT failure so the client can land on login cleanly.
} catch (err) {
+ res.clearCookie('token');
res.status(401).json({ error: 'Unauthorized' });
}Per-query in-memory similarity blocked the Node event loop as file count grew.
O(N) scans over 1,536-dim embeddings for every workspace query.
Filter by workspace at the DB layer first, then run similarity only on the active set.
- const allFiles = await File.find({});
+ const workspaceFiles = await File.find({
+ workspaceId: req.params.workspaceId,
+ deletedAt: null
+ });Post-mortems, system design notes, and engineering breakdowns from shipping real products.
An in-depth analysis of in-memory vector calculations for semantic search. Learn how V8 memory heap limits break calculations at 1,000,000 files, and how to scale dynamically using scoped MongoDB filtering.
Debugging expired tokens in Next.js applications. A deep dive into why checking cookie existence in middleware leads to infinite redirect loops, and how to safely synchronize state via backend headers.
A practical guide to structuring async FastAPI services with SQLAlchemy 2.0, connection pooling, and request-scoped sessions — patterns I use in production backends.
Multi-stage builds, non-root users, health checks, and env validation — the Docker habits that prevent painful production surprises.
From static generation and SEO metadata to performance hygiene and deploy config — lessons from shipping my own engineering portfolio on Vercel.
Frontend, backend, data, and cloud — the tools behind production systems.
Verified cloud expertise and peer-reviewed research — proof beyond the portfolio.
Certification
Amazon Web Services · Credly verified
Cloud fundamentals, core AWS services, security, and cost-aware architecture.
Publication
IEEE ICCSAI 2025 · IEEE Xplore
Deep learning face detection under low light with comparative denoising pipelines on benchmark datasets.
Prefer long-form writeups? Read the blog →
Have a project or idea? Send a note — I usually reply within a day.
Open to freelance, full-time, and interesting collaborations around AI, backends, and product engineering.