AI Applications & Prompt Engineering
Designing and prompting LLM systems that hold up outside the demo.
Transitioning into tech & analytics · open to analyst, data & AI roles
I build AI applications, data pipelines, and analytics.
I'm Tommy Yann. I spent years as a top sales rep; this past year I've been building instead — a multi-agent orchestration platform, a self-hosted LLM gateway, MCP servers, and a few full-stack apps, mostly in Python. Now I'm moving into tech and analytics for good.
4Projects shipped this year
256KToken context benchmarked
110–120%Of sales target · 10+ KPIs
About
From closing deals to shipping code.
I'm transitioning from sales into tech and analytics. Over the past year I've designed and shipped real systems on my own — the kind listed under Work below — while completing a B.S. in Artificial Intelligence at UMGC. I'm a lifelong learner, comfortable across Windows, macOS, and Linux, and open to analyst, data, and AI roles. Here's where I focus.
Automation & Workflows
Wiring tools into automations and agentic workflows with n8n and LangGraph.
Data Analysis & Engineering
Building pipelines that collect, clean, and move data so it's actually usable.
End-to-end shipping
Scoping a problem, connecting the pieces with these tools, and getting to something that actually runs.
Why me
What I bring.
What I've built this year, the stack I built it with, and what sales taught me.
Proven work ethic
Years as a top sales performer. The habit that transfers: I show up and hit the number.
A modern AI stack
Hands-on across the modern stack: Python, Docker, FastAPI, LiteLLM, MCP servers, and multi-agent frameworks.
Consistent output
Consistently 110–120% of target across 10+ KPIs in sales. I treat this field the same way: showing up daily and shipping.
Selected work
What I've built.
Real systems I've designed and shipped over the past year — self-directed, end to end.
01
AI agent orchestration platform
Containerized team-based agents with guardrail hooks, automated run verification, resource governance, and trust-tier checks so agents can't self-certify unverified output.
02
Self-hosted LLM gateway
Unifies local Ollama models and commercial LLMs behind one API with fallback routing, master-key auth, and a CLI-to-API shim — benchmarked up to 256K-token context.
03
AI agent security & guardrails
Sandboxed egress with proxy-based allowlisting, prompt-injection defenses, credential-leak scanning, and a stage/curate/promote intake pipeline for agent output.
04
Full-stack AI apps & pipelines
Shipped apps on FastAPI, SQLite, Docker, and Cloudflare — including automated payment reconciliation over SMS — plus MCP servers and a multi-tenant AI site deploy pipeline.
Now
Foundations & credentials.
The degree and certs behind the projects.
- B.S. in Artificial Intelligence University of Maryland Global Campus · AI Applications track · in progress Now
- Certifications in AI & automation Prompt engineering, AI applications, and data — underway Now
- GitHub portfolio Workflows, scraping & pipelines — building in the open Now
Background
Where I'm coming from.
Now
AI Applications & Prompt Engineering
Self-directed
Built a multi-agent orchestration platform, a self-hosted LLM gateway, MCP servers, agent security, and full-stack apps (Python, Docker, FastAPI, LiteLLM). See Work for details.
Sales background
Wireless Sales Specialist
Retail · B2B · Telesales
Top performer across all product lines; President's Club & Winner's Circle; consistently 110–120% across 10+ tracked KPIs. That discipline is what I'm carrying into tech and analytics.
In progress
B.S. in Artificial Intelligence
University of Maryland Global Campus · AI Applications track
Coursework in AI, data, and applied software.
Contact
[email protected]
Get in touch.
I'm looking for a full-time role in analytics, data, or AI, and open to freelance in the meantime. If you're hiring, I'd like to hear from you.