GTM Engineer · San Francisco Bay Area

I turn GTM operations into pipeline.

Sales, RevOps, automation, and software in one seat. I find what is slowing revenue down on the front line, turn it into technical requirements, ship the system, and measure what changes.

UC Santa Barbara · TypeScript · SQL/Postgres · APIs · Webhooks · Row-Level Security

Daniel Gonzalez
Daniel GonzalezGTM Engineer · UCSB ’26 · LinkedIn ↗
Results, scope & workflow design

What the systems changed.

~$10KWritten premium influencedOwner-reported pilot result across policies closed by two producers. Written premium, not revenue; measurement period and attribution breakdown are not published.
2–3 wks → same dayLoad Desk cycle timeA design comparison, not a measured field result: an outsourced Manual-J process redesigned into a same-day in-app workflow, demonstrable end to end in the open sandbox.
88Checkpoints made verifiablePhysical facility work a supervisor can now confirm digitally, in production for the staffing agency whose crew works an Uber site.

Referral Desk: 14 public helper tests are runnable. The separate private suite was reported as 1,666 checks; its run and deployment gate are not established by the public CI result. Every number’s source →

Flagship · Pilot · Private production data

Referral Desk

Producers at a Farmers agency found listing activity too late and had no structured follow-up or attribution. Referral Desk watches the signals, qualifies and routes them, and enforces the compliance rules in the database itself.

Business

Late signal discovery, unstructured follow-up, no revenue attribution.

Engineering

Postgres-enforced compliance: RLS tenant isolation, pre-send trigger gate, append-only outreach log. 14 public helper tests; a separate private suite reported as 1,666 checks.

Use

Piloted at the agency; producers worked real referrals on live client data.

Outcome

~$10K written premium influenced across policies closed by two producers.

Private production system. It runs on a live agency’s encrypted client data, so there is no public login. The architecture is public, and I will walk through the code and database live. Read the full technical case →  ·  Inspect the code on GitHub →  ·  Request a walkthrough →

Production · Private

ProWAX

Built for the staffing agency whose crew works an Uber facility. The crew taps NFC tags at 88 checkpoints, so supervisors see what was done, where, and when.

Business

Task completion was self-reported; supervisors could not verify work happened.

Engineering

NFC scan flow, tamper-flagged event history, supervisor console, security-hardened PWA.

Use

In production, used on shift by the agency’s crew at the Uber site.

Outcome

88 checkpoints digitally verifiable.

Working system · Approved for use

Classroom Narrative AI

Turns classroom evidence into report-card narratives without the school handing over student names. Every student gets a random alias, and the map from that alias back to the real name is designed to stay on the teacher’s own device. Narrative content remains pseudonymous and may still identify a student.

Business

Report-card comments cost teachers days, and general AI tools want the one thing a teacher must not hand over: who the student is.

Engineering

Random per-student aliases, row-level security per school, an append-only audit log, a private test count reported as 336 checks across 26 files.

Use

Approved for use in a Mountain View school district.

Outcome

The public demo opens as Guest in No Save mode. Use its built-in sample student.

Read the technical case →  ·  Open the demo →

Open demos

Open these yourself.

Three you can open right now.

Real problems, inspectable evidence. See everything →

Daniel with friends in UC Santa Barbara graduation gowns and blue sashes, arms around each other, on a bluff above the Pacific.
UCSB graduation with my friends, 2026.
A little about me

I sell insurance and I build software.

I sold devices at a Verizon authorized retailer in Goleta and grew accounts through service plans and upgrades. Now I sell insurance at a Farmers agency. I hold P&C and Life & Health licenses. I have used Salesforce, Outreach, and HubSpot as a rep.

I also build. I built Referral Desk and piloted it at a Farmers agency. Its rules live in Postgres, so a suppressed contact or an opt-out is blocked before anything sends. Classroom Narrative AI is approved for use in a Mountain View school district.

For Load Desk I redesigned an outsourced Manual-J process so it can run the same day inside the app. On every project on this page I wrote the spec, built it, and delivered it. The demos above are open, so try one. I still want to learn and get better at this.

San Francisco Bay Area

I look forward to connecting.

Open to GTM Engineer, Growth Engineer, and Revenue Operations roles in the Bay Area. I want a GTM team that knows how to scale. Each project is labelled with its real maturity on the Work page.

In twenty minutes I can walk you through Referral Desk end to end: the data model, the routing logic, the compliance rules enforced in Postgres, what broke during the pilot, and what I would build differently now.

Email me LinkedIn ↗ Resume →
Tell me straight

What did I miss?

If you hire for GTM, you see this page from an angle I cannot. I’d love to hear from you.

Feedback is how I get better. If you see something wrong, let me know.

Sends straight to my inbox. Nothing is stored on this page.