Jamison Miles

Senior Software Engineer · Austin, TX

Jamison Anthony Miles

I build fast, dependable software — from data pipelines that scale to half a million entities at a time to product features that move revenue.

Explore my experienceGet in touch
years of professional experience
6+
list capacity scaled at Dun & Bradstreet
50×
daily ad revenue at Atmosphere TV
+5%

About me

Background

Computer science by training, product-minded by practice. I care about systems that hold up under real load and features that get used by real people.

I'm a senior software engineer with experience across the stack — scaling data pipelines at Dun & Bradstreet, launching products and leading ad-tech integrations at Atmosphere TV, and building compliance and supply-chain tooling at Dell Technologies. My work usually lives somewhere between "how do we take a job that runs for days and finish it in hours" and "how do we make this feature effortless for the person using it."

Education

B.S. in Computer Science

University of Texas at Austin · 2016 – 2020

  • GPA 3.5 / 4.0
  • Freshman Research Initiative — Autonomous Intelligent Robots stream

Core stack

TypeScript / JavaScriptJavaC / C++Next.jsReactVueNestJSPHP (Laravel)ScalaAWS

Also worked with

PythonClojureGoogle CloudAzure

Where I've worked

Experience

From battery-powered trucks to ad-tech to enterprise data platforms — each stop sharpened how I build, ship, and scale.

  1. Dun & Bradstreet

    Jul 2024 – Present

    Senior Software Engineer

    • Re-architected the list upload system from synchronous to asynchronous, scaling capacity from 10K to 500K entities per list without losing performance under heavy load.
    • Refactored the internal record-update pipeline, cutting processing time from days to hours.
    • Co-led the integration of healthcare data into the Hoovers platform — designed the data-processing pipeline and built the APIs that serve it.
    • Kept API contracts up to date in close partnership with frontend developers, reducing QA overhead.

  2. Atmosphere TV

    Oct 2022 – May 2024

    Software Engineer → Software Engineer II

    • Shipped Shoutouts — a photo-and-caption upload product for venue displays — end to end in four months using NestJS APIs and a Vue web app.
    • Served as technical lead for integrating Google Ads Manager into the ad-fetching system, adding ~$1,000/day in revenue and positioning inventory for the Ad Exchange market.
    • Improved the ad-proxy system for fetching and serving, contributing logic that boosted daily ad revenue by ~5%.
    • Implemented the Ware2Go API, cutting device shipment time from 7–10 business days to 2–3.

  3. Dell Technologies

    Aug 2020 – Oct 2022

    Software Engineer I → II · Digital Supply Chain

    • Implemented rules and debugging for an audit engine built on Scala and Akka.
    • Built an internal Python + TensorFlow tooling project used by management to assess failures in cloud-based systems.
    • Delivered an internal Angular web app for sales management to view and monitor Global Trade Profiles, backed by Spring Boot REST APIs.

  4. Hyliion

    May – Oct 2018

    Software Engineering Intern

    • Built C, C++, and JavaScript tooling that pushed diagnostic data from hybrid truck systems to a web page for quick component triage.
    • Created scripts, code, and documentation for software installation, diagnostic execution, and performance analysis of battery-powered trucks.

Personal projects

Things I've built on my own time

Side projects, experiments, and open-source work — the places where I get to choose my own stack and my own problem.

gen3d-forge

Local, batch pipeline for game-ready organic props — a text prompt or reference image goes in, and scaled, LOD'd, PBR-textured GLB/FBX comes out. Characters get rigged (walk + idle) and any asset can be pixelated into sprite sheets for Godot. Everything runs locally on one GPU; an MCP server lets an LLM agent drive the whole loop.

TRELLIS.2FLUX.1-schnellBlender (headless)UniRigMCPGodot
gen3d-forge pipeline architecture

localCoderMcp

MCP server that lets Claude Code hand well-scoped implementation work to a local model on Ollama, so Claude plans and reviews while my own GPU writes the code. Each task runs in an isolated git worktree; the server reads the context files, runs the tests and feeds failures back to the model itself, then returns a compact diff for Claude to accept, revise or discard. Nothing touches the real working tree until it's accepted.

PythonMCPOllamaQwen3.8git worktrees
localCoderMcp request flow between Claude Code, the server, Ollama and the repo

Contact

Let's work together

Whether it's a role, a project, or a hard problem worth talking through — my inbox is open. References available upon request.

Email me

jamisonmiles234@gmail.comSend an email