I build AI agents for sales and support teams.
I design, ship and run them in production, along with the data models and evals underneath.
01 / WHAT I BUILD
Agents and automation for sales and support
The kind of work I do. The diagram shows the general shape of an outbound agent.
Outbound prospecting
Agents that research an account, find the right contacts and write a personal first email. I also run the sending setup behind them, so the emails land.
Call-to-CRM automation
An assistant that turns sales calls into CRM updates, so reps don't have to write up every call by hand.
Support agents
Agents that resolve routine tickets and hand the rest to a person. I review those handoffs to decide what the agents learn next.
Data pipelines & reporting
Data models and dashboards that give the team and the agents one set of numbers to trust.
Evals & observability
Every run is traced. Failures become test cases, and changes pass evals before they ship.
Agent cost & ROI
I track what each agent costs against what it produces, and cut spend with prompt caching and the right model for each step.
02 / HOW I WORK
How I build
Tie it to a number
Every process I build answers to a business metric: pipeline, time saved or data quality.
Prompting and context first
Proper prompting and context are the key to agentic success. Each agent gets clear instructions and the exact data it needs for its job.
Report on every step
I own the reporting on every process I build, down to each step of an agent's flow, so there's clear visibility into accuracy, efficiency and gaps.
03 / EXPERIENCE
From analytics to applied AI
I joined Zello as a Business Data Analyst and found myself spending more and more time building AI agents and automation. At some point, the old title just stopped being accurate.
- 2026 → now
Applied AI Analyst II · Zello
AI agents and automation for sales and support, plus the data, evals and reporting behind them.
- 2024 → 2026
Business Data Analyst · Zello
Analytics and BI across the business, then promoted into applied AI in 2026.
- 2022 → 2023
Data Analyst · Entrio
ETL, web scraping and Tableau reporting. I also managed our group of interns.
- 2020 → 2022
Data Analyst · Navat Capital Management
Investment and KPI data in R and Python.
04 / STACK
Tools I use
Agents & LLMs
LangGraph
LangChain
LangSmith
Claude · OpenAI
MCP
Data
SQL
Snowflake
dbt
Python · pandas
R
GTM and BI
HubSpot
Windmill
Gong
Omni · Hex
Tableau
Building for fun
Godot 4
GDScript
Blender
Claude Code · Codex
Git
05 / PASSION PROJECT
Worldforge
An original science-fantasy RTS I'm building in Godot 4, with a world and a novel to go with it. It's also where I test multi-agent engineering ideas before using them at work.
What's in it so far
Squads fight in formation, and the enemy AI runs its own economy.
- 4 factions, 65 unit chassis across 3 tiers, heroes, tech trees and siege
- AI-vs-AI match simulator for balance testing
- Built by a multi-agent team: Claude Code, Codex and Gemini working in one repo and reviewing each other
