John Steinmeyer

AI Systems Architect & Creative Technologist

I design and build the custom AI systems that bridge ambitious creative vision and powerful technical execution. With a deep foundation in machine learning and extensive experience pioneering the generative AI revolution, I partner with creative agencies and tech leaders to move beyond off-the-shelf tools and create scalable, production-ready AI pipelines for everything from national ad campaigns to enterprise-level automation.

Selected Works & Case Studies

Enterprise AI Automation for a AAA Gaming Studio

Client: Major MMORPG Publisher (under NDA)

Architected and deployed a suite of bespoke AI systems to solve critical business challenges, automating workflows, accelerating developer productivity, and creating a novel, data-driven executive reporting pipeline.

  • RAG-Powered Developer Assistant: Engineered a custom RAG agent on Discord, giving developers instant, context-aware answers from technical documentation and live SQL database queries, significantly reducing research time.
  • AI-Powered Bug Triage System: Developed an agent to automatically analyze Trello bug reports, perform diagnostic SQL checks, and provide initial analysis, giving developers a critical head-start on resolving issues.
  • Multi-modal Executive Reporting: Built a pioneering system using Vision LLMs to analyze thousands of daily developer screenshots, auto-generating concise progress summaries for leadership and providing unparalleled insight into company-wide efforts.

Technologies: Python, RAG, FAISS, Multi-modal LLMs (Vision), SQL Integration, API Design, Trello API, Discord.js

Pioneering Generative Video for GoFundMe's National Campaign

Client: GoFundMe (via ArtClass Post-Production)

As the lead AI Technologist, I developed the core generative pipeline for a high-profile national commercial. My novel techniques for achieving temporal consistency in video resulted in a visually stunning campaign with nearly a million views.

  • Production Pipeline Architecture: Designed and deployed the complete Stable Diffusion pipeline on production RTX 3090s, including server setup, CUDA libraries, and remote access, creating an efficient, scalable environment for high-volume asset generation.
  • Creative R&D and Look Development: Led the R&D phase with the director to establish a unique visual style, pioneering techniques that pushed the boundaries of what was possible with AI video at the time.
  • Complex AI-VFX Execution: Executed intricate, frame-by-frame inpainting for key sequences (e.g., rebuilding a building) and leveraged Deforum for dynamic, fully-generated animated transitions that were critical to the campaign's narrative.

Technologies: Stable Diffusion, Python, CUDA, Inpainting Models, Deforum, RunwayML, SDXL, LoRA

AI-Driven Creative Campaigns & R&D

Client: ArtClass Post-Production, NotCo

Served as an ongoing AI partner for a top-tier post-production house, providing end-to-end solutions for various campaigns and internal projects that required deep technical and creative expertise.

  • Brand & Style Replication: Trained custom fine-tuned models and LoRAs on artist-specific work to generate new assets that perfectly matched their unique, established style.
  • Advanced Animation Workflows (NotCo): Developed a novel workflow combining Google's FILM interpolation with Stable Diffusion to create seamless "age progression" animations for a campaign highlighting animal welfare.
  • Consulting & Team Enablement: Provided direct training and instruction to creative team members, empowering them to effectively leverage generative tools in their own workflows.

Technologies: Model Fine-tuning, LoRA Training, Google FILM, SDXL, API Prototyping (OpenAI), Prompt Engineering

Prototyping an Autonomous LLM Game Agent

Client: Personal R&D

To explore the frontiers of agentic AI, I architected a system allowing an LLM to fully control a character in a complex MMORPG. The demo showcasing its ability to converse, navigate, and act in a live environment generated significant industry attention and led directly to my recruitment by a AAA game publisher.

  • Custom API Bridge: Engineered a novel Python workflow to bridge OpenAI's API with the game's proprietary scripting engine, enabling real-time, two-way communication between the LLM and the game world.
  • Agentic Tool-Use Framework: Designed a robust system prompt and tool framework that empowered the LLM to perceive its environment (`Script: Nearby`), interact with players, and execute complex in-game actions based on natural language.

Technologies: Python, OpenAI API, LLM System Prompting, Agentic AI Design, API Integration

My Technical Toolkit

I stay on the cutting edge of AI development, leveraging a wide range of state-of-the-art models and platforms to build the best possible solution for any given problem.

LLMs & Agents

  • OpenAI Models (GPT series)
  • Google Gemini Models (2.5 Pro)
  • Anthropic Claude Models
  • Open Source (Llama, Mixtral, etc.)
  • RAG & Vector Databases (FAISS)
  • Fine-Tuning & LoRA Adaptation

Image & Video Generation

  • Google (Nano Banana, Veo3)
  • Stable Diffusion (incl. SDXL, Deforum)
  • Open Source (Wan 2.2, Qwen)
  • Image to 3D (Hunyuan, Tripo)
  • Interpolation (FILM)
  • ControlNet & Inpainting

Audio & Supporting Tech

  • Text-to-Speech (ElevenLabs, OSS)
  • Music Generation (Suno)
  • Transcription & Diarization (Whisper)
  • Voice Cloning & Transformation
  • Core Stack (Python, CUDA, APIs)
  • Game Dev (Roblox, LUA)

How I Can Help

Custom AI Systems & Pipelines

For organizations ready to move beyond generic tools. I design and build scalable, proprietary AI systems tailored to your specific creative or operational needs.

  • Custom Generative Video Engines
  • Enterprise Automation Agents
  • At-Scale Image Processing
  • Bespoke RAG & Knowledge Systems

Generative AI for Creative Production

A high-level technical and creative partner for post-production houses and creative agencies. I execute complex AI-driven visuals and solve novel production challenges.

  • Advanced R&D for Look Development
  • Character & Brand Consistency (LoRA)
  • Complex AI-VFX & Transformations
  • Production Workflow Integration

Strategic AI Roadmapping & Prototyping

The essential first step for any ambitious AI project. A dedicated engagement to translate business goals into a clear, actionable technical blueprint, de-risking your investment.

  • Technical Feasibility Analysis
  • System Architecture Design
  • Phased Development Roadmapping
  • Proof-of-Concept Prototyping

About Me

Long before the generative AI boom, my journey began with a fundamental curiosity: could I teach a machine to perceive and act in a complex world? I built my first machine learning systems from scratch—image classifiers that learned to play classic video games like Ninja Gaiden by watching me play. This foundational work taught me that AI is about building systems that can interpret, decide, and execute.

When the generative revolution arrived with models like Stable Diffusion, I was ready. As an early experimenter, I developed novel techniques for coherent AI video, work that caught the eye of the creative industry and led to my collaboration with ArtClass on a national GoFundMe campaign.

That project solidified my approach as an architect of AI solutions. My work since has spanned from designing enterprise-grade automation for AAA gaming studios to developing cutting-edge VFX workflows for creative agencies. I thrive on deconstructing ambitious goals—whether creative or operational—and engineering the robust, scalable AI systems required to achieve them. From first principles to final production, I build the bridge between vision and reality.

Let's Build Something New.

I'm currently available for new projects and strategic consultations. If you have a challenge you believe AI can solve, or a creative vision that seems impossible, I would love to hear about it.