travissmith
travissmith / README.md

travis@smith

OS
Windows 11 · local-first
Host
Exigere Solutions · Founder
Kernel
agentic-systems architect
Uptime
8 yrs clinical documentation
Shell
PowerShell · bash
IDE
Claude Code · Codex CLI
Languages
Programming
Python · TypeScript · JavaScript · Rust
Data
SQL · PostgreSQL
Real
English
Agentic
Orchestration
LangGraph · custom agent loops
Patterns
RAG · multi-agent · tool-use
Protocol
MCP servers · structured output
Models
Local LLM
Qwen2.5-Coder · Llama 3.1
Speech
Whisper · local transcription
Embeddings
nomic-embed-text
Frontier
Provider-agnostic model interfaces
Runtime
Ollama · Transformers
Vector
Dense search · Rust vector index
Stack
Backend
FastAPI · PostgreSQL · Docker
Frontend
React · Next.js · Astro
Cloud
AWS · Microsoft Azure
Focus
Design
privacy-by-design
Practice
adversarial evals · memory
Contact
Location
Nashville, TN
Stats
AI modules
14 · implementation + evaluation
Systems
13 · source-reviewed cases
Prototypes
13 extensions · 48 vertical packages
Status
available for work

travis@smith:~$

About

I build AI applications that connect models, documents and business workflows: retrieval systems, MCP servers, structured output and working product prototypes.

Founder of Exigere Solutions: medical transcription, documentation processing, report preparation and quality assurance.

I own requirements, architecture, integration and evaluation, using AI-assisted development to move from an idea to a working prototype. I am proficient in AWS and Microsoft Azure. My Private AI Lab runs open-source models on owned workstation hardware.

Skills & tech

  • agentic-systems
  • multi-agent-orchestration
  • rag
  • memory
  • evaluation
  • python
  • typescript
  • rust
  • sql
  • local-models
  • mcp-servers
  • structured-output
  • fine-tuning
  • product-prototyping
  • AWS
  • Microsoft Azure

Projects

Connected local Whisper transcription, model scoring and searchable transcript turns in a FastAPI application, with two authored demonstration cases.

status Working prototype
  • whisper
  • local-models
  • retrieval
  • evidence
Python README

Built serial telemetry parsing, an asynchronous WebSocket dashboard and an Ollama copilot that interprets telemetry snapshots. Includes a synthetic replay demonstration.

status Working prototype
  • serial
  • telemetry
  • local-models
  • replay
Python README

Implemented SQLite-backed memory, scoped lexical retrieval and relationship expansion, with interfaces for multimodal context and local synthesis.

status Research prototype
  • memory
  • sqlite
  • scoped-retrieval
Python README
Etzira Website Showcase

Built a Next.js service catalog, guided concierge and interactive browser demonstrations that explain practical business workflows.

status Working prototype
  • nextjs
  • workflow-demos
  • product-design
TypeScript README
Private AI Lab Showcase

Launched and operating on an RTX 6000 Pro and Threadripper Pro 9965WX, using open-source models for private, on-device inference and local processing.

status Local deployment
  • local-inference
  • open-source-models
  • private-processing
Local infrastructure README

Built typed evidence models, registry corpus tools and a LangGraph workflow for discrepancy review, evaluated with constructed corruption cases.

status Research prototype
  • evidence-contracts
  • langgraph
  • evaluation
Python README