conduithq.it.com
I rebuild messy Salesforce and HubSpot orgs so production AI agents can actually run on them — clean data, reliable structure, real write-backs.
Most people can clean a CRM or ship an AI demo. I do both, and I make them talk to each other.
Currently building GTM systems at CModel Data, Inc.
About
I grew up in New Orleans. At some point I got tired of waiting for life to come to me, so I packed up and moved to the Bay Area — first one in my circle to do it. My whole family eventually followed. That's probably the most Troy thing I've ever done.
I graduated from San Jose State straight into the strangest possible moment. No ceremony, no real transition. My degree was in Organizational Studies, not CS or engineering. I learned the technical side by building things until they worked. I figured out the rest as I went.
The instinct that drives my work — noticing what's broken and fixing it — didn't start with any system or org. It started with me. I turned it on myself first. Then I realized I could point it at anything, and it worked the same way. I spent months charting live basketball games for a data analytics firm — tracking plays, verifying data, catching errors in real time. Then I built Courtside. That's usually how it goes.
The midnight builds aren't about hustle culture. I'm not trying to impress anyone with my hours. I'm building toward something real — a stable life for me and my wife. A garden. Farmer's market Saturdays. Barbecue competitions on the weekends. Enough freedom that work is something I choose, not something I have to do.
Conduit is the vehicle. The name means something — a strong, powerful connection between two points. Every system I ship, every tool I build, every problem I solve before anyone asks — it's all pointed at the same thing.
From New Orleans · Based in the Bay Area
Problems I Solve
The six patterns that keep breaking GTM teams — and what I actually do about them
Reps don't trust the pipeline. Managers override the forecast. Execs add a gut-check discount. The architecture was never designed for accuracy. I rebuild the CRM structures, field logic, and governance so the numbers finally mean something.
Not a prototype. I shipped a production AI agent (Claude + Python + Flask + CRM API) that scores, enriches, and writes back real records in real workflows. Repetitive research and scoring work gets absorbed — permanently.
If leadership is deciding from gut feel, spreadsheet exports, or reports that need a decoder ring, the dashboard layer is the problem. I've built six executive dashboards that went straight into board-level planning. Clear signal. No interpretation required.
HubSpot says one thing. Salesforce says another. The real answer lives in a spreadsheet. I build the integrations, field mappings, sync logic, and documentation so the gap closes and stays closed.
If only one person understands the system, it's not a system — it's a liability with a name attached. I build with documentation and guardrails so the team can actually use what gets shipped, not just the person who shipped it.
Ask what the team shipped last quarter and you get a shrug. The work happened — nobody recorded it. I keep the receipts as I build: change logs, before-and-after numbers, and case studies with real results. The impact is provable.
Professional Work
Selected work at CModel Data, Inc
Case Study 01
CModel had hundreds of high-value EDOs in HubSpot, and every record looked the same. I built an end-to-end pipeline — HubSpot list in, live web signal enrichment, Claude ICP scoring, automated write-back — so every scored company now carries a 0–100 fit score, a lane, and a rep-ready signal note.
Read Case StudyCase Study 03
Started as a broken email sequence. Diagnosed the root cause, expanded scope after spotting a larger opportunity, and built an 8-workflow automation system across four platforms. Lead gen brought in-house. Outsourced vendor replaced.
Read Case StudyCase Study 10
Leads landed in Salesforce the same way whether the company was a total stranger or a five-year customer — nothing distinguished the two at creation. Built a Salesforce Flow that turns a lead's email domain into a verifiable match against every existing Account, routing real matches to a RevOps review queue and stamping everything else "Net New – Verified." Caught a blank-email edge case and an environment-specific queue ID, both before they could reach production.
Read Case StudyCase Study 12
A same-day Slack request to build a Services pipeline turned into a two-day architecture build — a new Sales Process, Record Type, required-field enforcement, and document-upload automation that turns checkboxes into proof instead of a claim. Along the way, a routine profile check surfaced something much bigger: the entire sales org had been defaulting to Salesforce's fallback Record Type for years, invisible because nothing ever looked broken. Flagged and fixed the same day it was found.
Read Case StudyPrior Engagement
A youth-focused nonprofit — migrating a $2.2M fundraising operation from Kindful to Salesforce
Prior Engagement · Case Study 01
A $2.2M nonprofit fundraising operation was running on Kindful and Google Sheets and had outgrown both. I sat in the architecture meetings from day one — had input on how the Salesforce org was structured — then built the full operational layer: repeatable processes, donor and event tracking, custom Apex automations, and six executive dashboards used directly in board-level planning. My first Salesforce engagement. No template. I built one.
Read Case StudyCore Skills
Think Tank
Personal builds. Real problems. Built solo.
AI Content Pipeline
Built a fully automated YouTube content pipeline. Scrapes trends, scores scripts against my content strategy, gates on human approval, generates voiceover, sources matching video clips from Pexels, and assembles the final video. Built for myself. Adopted by my employer. Growing channel with 1.5K weekly views.
Watch on YouTubeFull Stack App
A relocation planning app with real cost calculations, housing delta comparisons, peak season data, and an auto-generated 28-task moving checklist with timelines. Built because spreadsheets aren't good enough for a decision this big.
View ProjectScout Intelligence Tool
A full-featured basketball prospect tracker built for scouts who take accuracy seriously. 44 data fields, 11 evaluation sliders, analytics dashboard, CSV import/export, and printable scouting reports. The differentiator: a bias mitigation system most tools don't think about — blind eval mode hides player identity while scoring, a second-opinion engine flags high-discrepancy evals, and extreme scores require written justification before saving. Built because scouts have blind spots. This catches them.
View ProjectMore builds in progress. Deal Hunter coming soon.
Contact
Currently building GTM systems at CModel Data, Inc.