Projects

Problem

Diagnosing Kubernetes incidents requires cross-referencing pod events, logs, node pressure, resource quotas, and deployment state — repetitive work that takes time and deep context every time something breaks.

Approach

Built a ReAct agent loop with LangGraph on top of Claude, backed by 18 diagnostic tools covering pods, logs, events, node conditions, deployments, quotas, and OpenShift-native resources. Mutation tools (restart, scale) pause for explicit human approval via LangGraph interrupt before any write executes. Multi-cluster routing via JWT claims with per-namespace RBAC enforcement. FastAPI streams the agent's reasoning as Server-Sent Events. Ships with a Helm chart, ArgoCD manifests, and a Prometheus/Grafana observability stack.

Outcome

A production-grade agent that can reason through CrashLoopBackOff, pending pods, quota exhaustion, and rollout failures — and act on them safely, with a full approval audit trail and built-in observability.

Problem

Debugging CI failures is slow and context-heavy. Engineers manually cross-reference logs, run history, and code changes just to tell a real regression from a flake.

Approach

Modular pipeline: fetch job logs, PR diff, and recent run history from the GitHub API; sanitize and truncate to token budget; run deterministic flakiness heuristics (same-SHA pass/fail, transition count in history); then call Claude with Zod-validated structured output for root cause and fix suggestion. PR comments are idempotent via a hidden HTML signature — subsequent failures update the existing comment rather than duplicating it.

Outcome

CI triage in seconds — the PR comment surfaces cause, suggested fix, and a flakiness confidence score before an engineer opens the logs.

Problem

A transport operation needed a reliable, secure ERP it could roll out progressively — without re-architecting at every stage from a developer's laptop to the cloud.

Approach

Designed a multi-environment path from a Rocky Linux VM through OpenShift CRC to Azure Red Hat OpenShift, automated application delivery with Argo CD (GitOps), and added observability with Grafana and Loki. Reverse proxy and load balancing, MFA, stateless services, and externally managed configuration and secrets.

Outcome

Shipped a repeatable deployment pipeline that promotes the same artifacts across environments, with production-grade reliability and security built in from day one.

Problem

DevOps is hard to learn from reading — it needs real clusters, real pipelines, and the chance to break and fix things safely.

Approach

Container-first design with a Go backend, Angular frontend, and PostgreSQL persistence. Scenario-based labs (including Kubernetes) give each learner an isolated sandbox that mirrors real CI/CD and runtime troubleshooting.

Outcome

Guided learning paths tied to executable environments, so skill-building happens by doing rather than watching.

Problem

The annual career forum ran on manual, error-prone coordination across company registration, booth allocation, and contracts.

Approach

Built a platform that automates company registration, booth booking, and contract workflows end-to-end, alongside the club's internal tooling.

Outcome

Streamlined the forum for 20+ participating companies and replaced manual coordination with an automated workflow.

Problem

Content teams had no objective, repeatable way to judge whether an article was optimized for search.

Approach

Built independently with Spring Boot, Next.js, and MongoDB. Article-scoring workflows combine web scraping, NLP, and OpenAI API signals into a practical optimization score.

Outcome

Turned fuzzy content-quality signals into concrete scores teams could act on.