Tanner Stevens

Senior Software Engineer · Full-stack

Full-stack engineer with a decade of experience delivering user-facing products end-to-end – from frontend component architecture and interactive UI to backend services, REST APIs, and data pipelines. Consistently embedded with Product and domain teams to ship complex initiatives across organizational boundaries, making specialized capabilities accessible to non-technical users.

Experience

Freelance Software Engineer

2026 – Present

  • Building a data bridge for an ISP/MSP client that keeps Odoo, Syncro, and Powercode in sync bidirectionally — a mesh topology where declared authorities and recency rules decide which system's version of a record wins, with idempotent operations that lean on the connected systems as source of truth and SQLite only for cached lookups. Data flows are authored at run time through an author → preview → publish loop with no deploy, exposed as an authoring API meant to be driven by an AI agent on the user's behalf
  • Extending the same bridge to GPSInsight fleet telemetry — compiling raw vehicle data into per-customer trips so mileage and time spent feed billing directly, and techs can review and confirm those same trips to generate timesheet data

The Aerospace Corporation

Jan. 2016 – 2026

  1. Engineering Specialist 2023 Huntsville, AL / Remote
  2. IR&D Portfolio Manager (role) 2021 – 2023 Remote
  3. Senior Member of Technical Staff 2020 El Segundo, CA / Huntsville, AL / Remote
  4. Member of Technical Staff 2018 El Segundo, CA
  5. Associate Member of Technical Staff 2017 El Segundo, CA
  6. Intern 2016 El Segundo, CA
  • Built a WYSIWYG site editor for non-technical staff where configurable widgets serialized to HTML stored in a database, then rehydrated back into live React components – supporting a draft/publish staging model so maintainers could edit and preview changes live before publishing to end users
  • Sole engineer on a cross-functional team that won a Corporate Innovation Award – converted a domain expert's MATLAB proof-of-concept into a configurable, repeatable production pipeline, operationalizing a novel cross-domain forecasting capability previously locked to a single specialist
  • Inherited a critical deliverable mid-cycle with no handoff — multiple deadlines already in flight across a four-part system whose only documented interface was tribal knowledge accumulated across three engineer generations. Reverse-engineered intended behavior from artifacts alone, unified disparate scripts behind a single metadata-driven entry point, and designed a base-class architecture that handled ongoing data format evolution without full rewrites
  • Mentored interns and peers through pair programming and hands-on problem decomposition – engaged with early career programs across a decade, from staffing intern events to working directly alongside junior engineers on complex technical challenges
  • Designed and implemented a self-describing service mesh — defined a machine-readable schema standard that each containerized service used to broadcast its own capabilities, enabling a single routing layer to dynamically dispatch REST requests across organizational boundaries without specialized knowledge of the underlying tools
  • Architected a multi-user real-time system on a pub-sub message broker – independent actors negotiating state via message-passing, each maintaining an asymmetric local view derived from message history, with UI clients implemented as first-class subscribers on the same bus – designed to scale to tens of thousands of concurrent actors through async event resolution and progressive broker sophistication
  • Built interactive 3D visualizations in ThreeJS and D3 — including a scenario planning tool that replaced physical artifacts in live operations and a multi-timeline astronomical renderer spanning multiple coordinate systems and reference frames
  • Stood up the organization's first internally hosted LLM platform serving hundreds of internal users — architected to handle sensitive data with no funding or access gates, offering a rotating selection of self-hosted open-source models that scaled to incorporate larger enterprise models as infrastructure grew, with full parameter exposure and RAG pipelines
  • Designed and built a document processing pipeline implementing LLM-based map-reduce: filter prompts run per chunk, summary prompts consolidate results, and a reduce prompt iterates until convergence — delivering staged artifacts and a compiled report to the user
  • Integrated an LLM into an existing semantic document analysis tool to automate keyword generation, removing a major domain specialist time bottleneck and lowering the domain knowledge barrier for general analysts

Skills

Languages
JavaScript, TypeScript, Python, C/C++, SQL, C#
Frontend
React, ThreeJS, D3, NextJS, Electron
Backend
Node.js, Express, Django
Infrastructure
Docker, Kubernetes, Linux, AWS, Woodpecker CI, Jenkins
Databases
MongoDB, Redis, MySQL, SQLite
AI / ML
LLMs, RAG, prompt engineering
Tools
Git, GitHub, Jira

Beyond Work

Education

Northern Arizona University — B.S. Computer Science, 2017 · Flagstaff, AZ