SWE @ ADP - Burned more money on my homelab than I ever would've on AWS. Would do it again.
Interested in how LLMs can augment data into the right format for optimization models.
JellyfinMCPServer — MCP server giving Claude natural-language access to a self-hosted Jellyfin media library.
Jellyfin Smart Search — Semantic video search over a self-hosted media library using SigLIP embeddings on distinct frames and nomic-embed-text on captions.
Homelabbing — Kubernetes Cluster w/ Talos, Proxmox cluster, TrueNAS, self-hosted Postgres/pgvector, *arr stack (iykyk).
In the pipeline — AI agents + optimization systems.
Languages — Go · Python · TypeScript · Java · SQL
AI / ML — PyTorch · HuggingFace · LangGraph · FastMCP · CVXPY · OR-Tools
Infra — AWS (EKS, ECR, S3, Route 53) · Kubernetes · Docker · Proxmox · GitHub Actions
Databases — PostgreSQL · MongoDB · Redis · ChromaDB · Apache Ignite