Backend & Applied AI Engineer
Building distributed systems, cloud control planes, and governed agentic workflows.
I build product-focused systems where backend architecture, infrastructure, and AI have to work together reliably.
My recent work covers:
- distributed and event-driven backend systems
- cloud control planes and runtime lifecycle orchestration
- Kubernetes, GitOps, CI/CD, and observability
- policy-grounded RAG, guarded agents, workflow memory, and evaluation
- human-in-the-loop AI with traceable and governed decisions
I care about the engineering around the model just as much as the model call: validation, failure handling, idempotency, observability, evaluation, safety boundaries, and operational proof.
A governed agentic workflow that turns an unstructured motor-insurance claim into a policy-grounded, human-reviewed, and fully traceable case.
- document extraction and deterministic validation
- policy RAG with verified citations and abstention
- guarded tool-calling agent and human review
- safe workflow memory from trusted outcomes
- AI gateway, per-run traces, cost and latency visibility
- synthetic evaluation suites covering the complete workflow
Built with: TypeScript, Next.js, PostgreSQL, Prisma, pgvector, OpenAI, Bun, Turborepo, Docker
A control-plane-first cloud workspace platform that turns a browser request into a real code-server environment running on AWS.
- EC2 Auto Scaling Group allocation and idle VM reuse
- explicit runtime lifecycle and failure states
- Redis locks and runtime state mirroring
- VM agent and Docker-based workspace boot
- S3-backed project restore and synchronization
- browser dashboard exposing infrastructure state
Built with: TypeScript, Next.js, PostgreSQL, Redis, AWS EC2/ASG/S3, Docker, Clerk, Bun, Turborepo
A production-style uptime monitoring platform built around independent workers and Redis Streams.
- scheduled website monitoring and response-time history
- status-transition detection and incident lifecycle
- asynchronous notification pipeline
- authenticated user and admin dashboards
- Prometheus metrics and containerized local development
- Kubernetes deployment managed through a separate GitOps repository
Built with: Go, Gin, Redis Streams, PostgreSQL, Next.js, Docker, Prometheus
Deployment: runstate-gitops — Argo CD, Kustomize, External Secrets, ingress, Grafana, HPA, and automated image updates.
Languages: Go, TypeScript, JavaScript, Python, SQL
Backend: Gin, Node.js, REST APIs, background workers, event-driven systems
Data: PostgreSQL, Prisma, Redis, pgvector
Cloud & infrastructure: AWS, Docker, Kubernetes, Argo CD, Kustomize, GitHub Actions
AI systems: RAG, tool-calling agents, workflow memory, evaluations, tracing, human review
Frontend: Next.js, React, Tailwind CSS
Observability: Prometheus, Grafana, structured traces and model-call metadata
I am interested in backend, platform, cloud, and applied AI engineering roles where I can work on reliable product systems—not isolated demos.
I am especially drawn to teams building developer infrastructure, AI-enabled workflows, distributed services, internal platforms, or operational tooling.
Explore the repositories above for architecture diagrams, implementation walkthroughs, demos, evaluation results, and operational evidence.