Ikenna “Navy”
Nwakwesi

Dia turns a client discovery questionnaire into a scored opportunity and three ready-to-send documents: client profile, opportunity analysis, and proposal draft.

Dual architecture: a fully deterministic scoring and document pipeline, plus a Gemini LlmAgent that uses those deterministic functions as tools and degrades gracefully without an API key. Ships with a dependency-free voice-enabled UI (TTS, dictation, hands-free mode), a pytest suite, an eval directory, a Dockerfile, and one-command Cloud Run deploy.

Run like a product, not a script: PM board, status directory, security policy, secret scanning, and 17 tracked issues.

Code: github.com/mdvnavy/Dia

An enterprise commercial dataset of 700,000+ transaction records, made queryable in plain English. Parquet columnar storage on an embedded DuckDB engine, Python and pandas ingestion with normalization and deduplication, a unified SQL view, and a Model Context Protocol server exposing it all as tools to an AI assistant.

Concept to verified product in roughly 18 hours on constrained hardware.

A 3-server VPS agent fleet on a private Tailscale mesh (WireGuard, ACLs, MagicDNS) running multiple agent runtimes per server, with self-served artifacts to mobile and integrated search and scraping services.

I connected a vendor’s long-term memory product over the tailnet before their integration guide existed, rejected their public-tunnel recommendation in favor of a zero-exposure architecture, and debugged an undocumented TLS SNI requirement along the way.

A distribution model for agent capabilities: skills packaged as versioned SKILL.md modules with runtime metadata, published to a Hugging Face dataset registry, and consumed by agents across the fleet.

Fourteen production skills covering the Google Workspace surface, plus a registry-of-registries index. Effectively a plugin model and developer on-ramp for agents.

Coming to code from design and data, you don’t see a developer’s tool - you see something with no ceiling: administration, organization, research, a dozen uses nobody thinks to apply it to. And on the moment it clicks: I can figure this out. No roadmap, just reading, researching, and pulling threads.

Notes from wiring a long-term memory service into a private mesh before the vendor published their integration guide. Their version suggested a public tunnel; mine never left the tailnet. A quiet case for private-by-default agent infrastructure.

Using your computer primarily through agents as the interface - you orchestrate the agent, and the agent navigates the system for you.