Amie Twyford

Growth Marketing & GTM Strategist
AI, Product & New Market Expansion

Me, Uncoded

I’m interested in how incentives influence human behavior. In how small shifts in structure can meaningfully change outcomes over time, and how choices compound beyond their initial moment, often in ways we don’t expect.Curiosity is how I make sense of the world. I like seeing how things work, how they grow, how they break, often and not always intentionally.

My Work, Outlined

I’ve led multi-million-dollar campaigns, helped early-stage businesses find traction and have translated data into measurable revenue.I focus on how systems and decision-making interact to drive repeatable outcomes.

Time, Well Spent

Outside of work, I spend time traveling, nonfiction reading and following ideas just a bit outside of my comfort zone. I've called Arkansas, Colorado, NYC and Mexico home.I’m interested in how people adapt to change, culturally, socially, and personally, and I tend to listen within my surroundings. I also write as a way to explore our world through a different set of glasses.

Back

What I Do

I’m interested in how users move through funnels, how behavior and intent correlate when both can be intangible, and how small shifts compound into meaningful outcomes over time.

Current FOcus

▪︎ Growth marketing (GTM strategy, funnels, experimentation)
▪︎ AI-driven user experience
▪︎ Product-led growth
▪︎ Red teaming

Back

projects + applied AI

🧷 A research intelligence system combining domain extraction, persistent RAG memory and conceptual drift detection across AI alignment literature. Analyzes cross-domain shifts weekly and traces epistemic roots across lab, independent and academic sources🧷 A multi-agent intelligence platform designed to track how the world's leading frontier models (ChatGPT, Claude, Gemini, and Perplexity) surface and conceptualize any custom vector🧷 Personal RAG system indexing research, projects and voice writing for semantic retrieval🧷 Agent observation platform, understanding workflows/audit logs outside of typical practices🧷 Evaluating reference pairs, identifying inversion rates, insight into reward modeling training; extending findings through lightweight fine-tuning experiments using Llama 3.21B, LoRA, Groq🧷 A multi-agent research system of four agents with distinct personalities and persistent memory, debating AI autonomy, observing emergent behaviors beyondTech Stack:
Python · n8n · ChromaDB · Pinecone ·Voyage AI · Supabase · Airtable · Notion · Streamlit · Claude · Anthropic API · FastAPI · Langsmith · LangGraph · Jupyter · DigitalOcean · Perplexity · Gemini API · Railway · Vercel · HuggingFace

Back

What I can't stop wondering

🫧 Will intelligent systems change when designed to listen, not respond
🫧 How people adapt after moments that force a reset
🫧 What systems learn when behavior is observed but not explained
🫧 Where assumption begins in how we measure intent
🫧 What it looks like when scale happens faster than the understanding
🫧 How responsibility will be preserved as systems outlast creators
🫧 What gets carried forward when contexts change faster than memory
🫧 How progress is measured when outcomes can take years to reveal