Resume.

Building and shipping applied-LLM products and infra with real adoption. RAG, knowledge graphs, agent systems.

• claude-memory-viz: embedding + clustering visualizer for Anthropic's Claude MCP memory (95★).
• xray: MCP for progressive code intelligence via ast-grep (43★).
• contextgraph: decision audit ledger for AI agents; full-stack SDK + server + UI.
• alpha: IAM policy rightsizing agent with AI risk signals and instant rollback.
• logsieve: log dedup sidecar using Drain3, production-ready Helm chart, Go.
• wiki-in-a-box: offline Wikipedia with hybrid no-index RAG.
• ita-kg: Income Tax Act knowledge graph + RAG for legal lookup.
• murmur: voice interface for Claude Code / Codex CLI via whisper.cpp with Metal accel.
• kubectl-smart: CLI that turns Kubernetes debugging into signal prioritization.

• Primary SRE for a tier-1 client; deployments, prod debugging, and maintenance.
• Led infra for a major POC; implemented Azure RBAC with managed identities and SDK-level deep dives.
• Cut AWS spend by ~$800K/yr as a core infra contributor on a cloud-agnostic migration.
• Elasticsearch snapshot archival saving ~$8,900/mo in storage.
• Multi-tenant monitoring with Grafana / Prometheus / Loki / Tempo, with retention strategy.
• Co-built ops automation platform in Go + Temporal for long-running maintenance workflows.
• Packaged Ansible as a Kubernetes Job via Helm for env bootstrapping (secrets, RBAC, base config).
• Mentored juniors; incident management within the team.

• Owned infra for a metrics+events ingest peaking at 350M events/min; tuned throughput, reliability, cost.
• AWS + GCP perf tuning and capacity planning; ~20% cost reduction via rightsizing.
• Guided Kubernetes orchestration choices; documented FMEA for critical paths.
• Streamlined onboarding via VPC peering and AWS PrivateLink; shipped metrics + log pipelines end-to-end.
• Hardened backups + security controls mapped to SOC 2; improved incident readiness.

• Scaled data pipelines to 10M+ packets/day; owned infra, observability, release pipelines.
• System design + interim tech lead + interim database engineer.
• Led HTTP → MQTT transition; significant cost and performance wins.
• MySQL → AWS RDS migration with near-zero downtime + DR plan.

AI infra and backend for a public-equity research SaaS.

• RAG over multi-source datasets with retrieval caching and query planning.

Delivered MVP during COVID.

• Flutter app + Python backend on AWS via Terraform; shipped to multiple restaurants in month one.

Telematics platform + cost-efficient location services.

• Scaled to ~20k devices, ~4M messages/day; self-hosted reverse geocoder cut infra ~80%.
• Owned frontend, backend, infra, database; led a small engineering team.
• Shipped products on Android, Rails, Node.js, Angular, React Native.
• Solved a complex timetabling problem with a genetic algorithm.