Professional Summary
Self-taught AI engineer with deep expertise across all modalities and a unique cross-domain background spanning banking/finance, technical support, and management.
Creator of the Multimodal Discord Bot, several Python/AI packages, and contributor to major open-source AI projects.
Proven ability to bridge the gap between cutting-edge AI research and practical business applications.
Key Differentiator: Combining deep AI technical expertise with real-world business acumen from banking, customer service, and operations management - enabling me to build AI solutions that actually solve business problems.
Key Projects & Contributions
Comprehensive AI assistant integrating text generation, image creation, music synthesis, voice generation, and web search.
Built with TypeScript and Docker, supporting hybrid local/cloud AI providers with 30+ commits and ongoing development.
TypeScript
Docker
Redis
PostgreSQL
Ollama
OpenRouter
Enhanced music generation system with structured input processing, Discord integration, and API overhaul
for integration with the Multimodal Discord Bot.
Features automatic chunking for Discord's 10MB limit and configurable output formats.
Python
PyTorch
FastAPI
Docker
Video generation framework optimization for practical deployment. Working with next-frame prediction neural networks
for progressive video generation with constant memory usage regardless of video length. API integration with
on-demand unloading for future usage with the Multimodal Discord Bot.
Python
PyTorch
Diffusion Models
Gradio
HuggingFace Model Quantization and TIES/DARE-TIES Model Merging
View Models
105+ quantized AI models including Qwen3 series, Llama family, DeepSeek models, and custom abliterated versions.
Specializing in GGUF quantization and model optimization for local deployment.
Experienced in merging and quantizing state‑of‑the‑art language models using techniques like
TIES and DARE‑TIES and exporting them to efficient formats for local inference. Adept at working
with transformer‑based LLMs, diffusion models, audio and video generation systems, and robotic/agentic pipelines.
GGML/GGUF
LLaMA.cpp
Quantization
Model Merge Strategies
Mergekit
Model Optimization
Qwen
DeepSeek
Kimi
Llama
Currently, three Python packages are published for AI development workflow: token counting for various file types,
file tree generation with icons, and requirements version checking against PyPI.
Python
Package Development
API Integration
AI Libraries
AI Workflows
Professional Background
Managing complex financial products, loan origination (NMLS 2071358), high-dollar transactions, and risk assessment.
Handling diverse account types, credit rehabilitation, and fraud mediation. This experience provides deep understanding
of regulatory compliance, risk management, and customer relationship building - crucial for AI applications in fintech.
Formerly licensed in 11 states, originated conventional, government, and jumbo loans. Developed expertise in credit analysis,
loan structuring, and regulatory compliance. Passed comprehensive SAFE exam demonstrating analytical and risk assessment capabilities.
Apple ACiT certified technician with Samsung Level 1 & 2 certifications. Performed device diagnostics, repairs,
data recovery, and insurance assessments. Developed technical troubleshooting skills and customer service expertise.
Managed 6 restaurants with 250+ employees across multiple locations. Responsible for operations, customer service,
staff training, and process optimization. Developed leadership, project management, and operational efficiency skills.