Skip to content
View TheJoshCode's full-sized avatar
🏠
Working from home
🏠
Working from home

Block or report TheJoshCode

Block user

Prevent this user from interacting with your repositories and sending you notifications. Learn more about blocking users.

You must be logged in to block users.

Content in all repositories owned by your account will be closed.
Maximum 250 characters. Please don’t include any personal information such as legal names or email addresses. This note will only be visible to you.
Report abuse

Contact GitHub support about this user’s behavior. Learn more about reporting abuse.

Report abuse
TheJoshCode/README.md

Joshua Roberts

Embedded AI Engineer | Edge Inference | Autonomous Systems

joshrobertsaiml@gmail.com · (812) 791-2800 · Paoli, Indiana, USA · github.com/TheJoshCode
U.S. Citizen · Clearance Eligible


Professional Summary

AI/ML engineer with 7 years building, shipping, and supporting real-time AI systems on constrained hardware. Python + C++ developer with CI/CD experience (GitHub Actions, Docker) automating builds and releases for Jetson-class edge deployments.

Co-founded Dream AI and led 10+ engineers shipping a real-time diffusion plugin to 100k+ users. Built Maveric AI, a voice agent on NVIDIA Jetson Orin NX with sub-300 ms end-to-end latency on live CAN bus telemetry — fielded with two racing teams and adapted for DoD field trials.


Core Competencies

Edge AI
NVIDIA Jetson (Orin NX) · PyTorch · ONNX Runtime · CUDA/GPU acceleration · real-time inference pipelines · edge containerization

Systems
C++ · Python · Linux · RTOS · ROS2 · CAN bus & vehicle telemetry · microcontroller & driver interfacing

Sensing & Integration
Sensor fusion (camera / LiDAR / IMU calibration) · HMI design


Professional Experience

Embedded AI & Hardware Engineer · Independent (stealth-stage startups)

Nov 2025 – Present

  • Engineer custom embedded AI inference engines, edge containerization setups, and 3D simulation tools for global clients.
  • Integrate low-latency telemetry pipelines and custom 3D asset workflows into real-time physical simulation tools.

Senior AI Engineer · Code19 Racing

Jan 2025 – Nov 2025

  • Built Maveric AI, an autonomous racing driver voice agent holding sub-300 ms end-to-end response latency on live CAN telemetry.
    Stack: Whisper large-v3-turbo (ASR) · Qwen 3 1.7B (LLM) · proprietary TTS · all-MiniLM-L6-v2 (RAG).
    Implemented in Python + C++ (llama.cpp bindings + inference code), containerized in a secure Docker image.
  • Built GitHub Actions CI/CD for fast, repeatable rebuilds of the Maveric Docker stack so fixes could ship to field hardware.
  • Fielded with Aipex and Rafa Racing across 3 events, including AI in The Fast Lane at Circuit of The Americas.
  • Transitioned Maveric into an autonomous voice agent drone system for DoD field trials.
  • Integrated telemetry with Anduril’s Lattice SDK for multi-agent drone coordination in simulated environments.
  • Led compute packaging and field deployment; supported DoD autonomous-agent trials (including Camp Atterbury TREX) using operator feedback for in-field adjustments.

Intern → Co-founder & Lead AI Engineer · Dream AI

May 2021 – Mar 2025

  • Co-built a real-time diffusion img2img filter plugin for Unreal Engine 4 on ONNX Runtime, reaching 100k+ users at peak.
  • Led a cross-functional team of 10+ AI, software, and 3D graphics engineers.
  • Partnered directly with NVIDIA and Epic Games engineering teams on custom spatial rendering pipelines, digital twin environments, and real-time asset generation models.
  • Built GitHub Actions CI/CD that automated builds and published directly to releases, enabling rapid edit-and-rebuild cycles for 100k+ users.

Education

Mechanical Engineering coursework (no degree conferred) · Indiana State University · Aug 2020 – May 2021
Kinematics · Dynamics · Statics · Systems Engineering

Dual-credit Computer Science Core 40 Technical Honors Certificate · Ivy Tech Community College · Aug 2018 – May 2020
Computer engineering basics · Python · C++ · Computer Literacy

Pinned Loading

  1. BlendNite BlendNite Public

    Blender Viewport Nanite GPU Clustering

    Python

  2. MU-TH-ER-2026-DEMO MU-TH-ER-2026-DEMO Public

    Alien MU/TH/ER 6000 Terminal in Unreal Engine 4, Running An Actual Offlline LLM As The Brain.

    Python 5

  3. OFFLINE-AI-BALL-KNOWLEDGE OFFLINE-AI-BALL-KNOWLEDGE Public

    My running stack list for on-device agents, with some advice for beginners.

    1