Jeremy Odam

Operations engineer turning AI into labor.

25 years in regulated utility operations. Currently building AI-augmented operational tooling for the field, the supervisor's desk, and everywhere in between. Pacific Northwest. Remote-first.

About

Twenty-five years inside a regulated utility. Field operations supervisor. Six AI certifications across Vanderbilt, IBM, and Google. Two consumer mobile apps in active beta on iOS and Android.

The combination is rare on purpose. Most operators do not learn to build. Most builders do not learn the operations floor. The people who do both end up with a different read on what AI is for.

What I am building is operational tooling that takes AI seriously as labor. Not a copilot. Not a chat widget. Real workflows that run alongside a 600-person field force, a 30-supervisor reporting org, and fleets under regulatory tool calibration. That is the work.

Selected Work

AI Field Operations Assistant

5 min → 30 sec

Conversational knowledge base for 600+ field operators in a Fortune-class regulated utility. Replaced 5+ minute manual lookups with 30-second natural-language queries. Citations surfaced with every answer, no hallucinated procedures. Pilot approved by enterprise leadership.

React. Claude API. RAG with native Citations. Vercel serverless.

Conversational BI in Microsoft Teams

92-95% time saved

Natural-language reporting layer over Power BI, embedded inside Microsoft Teams. Compressed routine 30-minute drag-and-drop sessions to 90-second prompts. Same data. Same governance. The bottleneck was the interface, not the data.

Claude API. Power BI semantic model. Teams app shell. Server-side query orchestration.

Vertical AI Platform

19 modules, 1 backend

Designed and built a multi-product AI platform spanning field ops, fleet, scheduling, safety, supervisor reporting, dispatch, command, and twelve more. Single universal Claude streaming endpoint. Config-driven multi-tenancy. Branded, scoped, and deployable per client in days.

Universal serverless dispatch. pgvector retrieval. Supabase RLS. Cross-tenant isolation in CI.

Operational Safety Assistant

Ships standard

Always-on safety assistant grounded in company procedures, NFPA 54, NFPA 58, and 49 CFR 192. Surfaces relevant procedures on demand. Routes to incident reporting when the conversation calls for it. Built on a single principle: the safer path should be the easier path.

Claude API. Native Citations. RAG over public safety standards. Lone-worker integration.

GarageBuddy AI

iOS + Android beta

Consumer DIY auto diagnostics app. Pairs with OBD-II hardware to diagnose vehicle issues, surface parts, and route to service. Capacitor-wrapped from the same web platform that powers the ops modules. Same architecture, native shell. Proves the platform extends from B2B SaaS to consumer mobile without rewriting the core.

Capacitor. iOS + Android native shells. Same Claude streaming backend.

Stack & Credentials

Stack

  • Claude API (Sonnet 4.5)
  • React, TypeScript, Next.js
  • Vercel serverless
  • Supabase, Postgres, pgvector
  • Capacitor (iOS, Android)
  • Python (RAG pipelines)

AI Certifications

  • Vanderbilt: Claude Code
  • Vanderbilt: AI Agents and Agentic AI
  • Vanderbilt: OpenAI GPTs
  • Vanderbilt: Generative AI (+1)
  • IBM: Prompt Engineering
  • Google: GenAI Leader

Operations

  • FEMA ICS-100, 200, 300
  • Regulated utility ops, 25 years
  • Field supervisor, 600+ operators
  • Multi-system platform integration

Open To

Remote senior IC and fractional roles in AI implementation, applied AI engineering, and ops automation. Forward Deployed Engineer, Applied AI Architect, Director of AI Implementation, Fractional COO at AI-native startups.