Senior Software EngineerI lead AI and platform work that spans teams, from design doc to production

Most recently: platform-wide permissions, AI features, reliability, and sales tooling at SewerAI.

Impact

Tackles ambiguity head on

Go-to-market · Sales

Built the sales demo platform

Built self-serve demo environments where prospects test the platform on their own data, with a one-step path from demo to live account. Demos built on the platform helped close ~$30M in ARR.

How I did it →
  • Prospects load and explore their own inspection data, which customers told us they valued most.

  • Demo accounts with user invites, so sales could run a demo without asking engineering.

  • A one-step flow that turns a demo into a live customer account, so a closed deal starts from what the prospect already saw.

  • Demo analytics, so sales could see which features prospects actually used.

Every customer account · Backend, workflows, UI

Redesigned how access works across the platform

Moved access from per-project grants to teams, matching how utilities and contractors organize crews. Approved design to production in 11 days; a pre-launch bug bash surfaced 10 issues before launch, including a privilege-escalation risk.

How I did it →
  • A user's access can come from several teams plus direct grants. I recorded where each grant came from, so removing one path recomputes access from the rest instead of wiping it.

  • Before migrating a large customer's ~1,300 projects, I dry-ran the plan against live data and caught 49 users the backfill would have hidden.

Product + engineering workflow

Put AI to work in the product, with users in control

I decided the product's AI assistant should take real actions, like restoring records or changing inspection types, but only after the user approves each change. That let it do more without a mistaken bulk edit ever reaching customer data.

How I did it →
  • Every write starts as a preview that lists what will change and what's blocked, with the reason, before the user approves.

  • Built the LLM project planner we launched publicly, and designed AI data mapping for professional services with adjustable levels of human review.

  • Ran a ~16-PR component-library migration largely through AI agents, with every PR human-reviewed.

Cross-service · No prior owner

Turned customer-found failures into a reliability program

Traced ~5 customer-reported incidents to one systemic cause, silent failures in background workflows, and started the fix before anyone asked.

How I did it →
  • Defined one job-status standard every workflow reports through, with stall detection based on the workflow engine's own state instead of tuned timeouts.

  • Flagged an incident with ~11,300 failed workflow runs, and closed a P1 ingestion incident in ~2 hours.

Company-wide support load · Product, CS, Engineering

Redirected the company's support-reduction effort

Analyzed 2,600+ support tickets: 22% weren't support work, 28% were repeats, and most top requests were product gaps. I dropped my own approved AI-agent plan and moved the project to the highest-volume fixes.

How I did it →
  • Two people handled 82% of tickets, and the existing bot resolved only ~11%. A better bot wouldn't fix missing features.

  • Every fix shipped with a tracking event and an in-app announcement, with a control group built into the measurement plan.

Experience

Where I've worked

SewerAI logo
AI inspection platform · Revenue-facing

SewerAI

Senior Software Engineer · 2025 – present

Shipped the truck features needed to onboard a new customer, taking truck seat subscriptions to 104% of target.

Emcee Studios logo
Company's core product · Full stack

Emcee Studios

Senior Software Engineer · 2023 – 2025

Took the company's social commerce app from zero to launch, owning it from architecture through release.

Figure logo
Multiple product lines · Shared platform

Figure

Software Engineer · 2021 – 2023

~80% faster private-label releases, +11% retention, and one unified dashboard replacing several.

PopSugar logo
Editorial + ad campaigns

PopSugar

Software Engineer III · 2019

Cut campaign development time 25% with reusable templates.

San Francisco International Airport logo
San Francisco International Airport logo
Airport-wide operations · Led engineers and vendors

San Francisco International Airport

Software Engineer · 2015 – 2019

Shipped the airport's first native app (~60% faster incident response), cut incident overhead 20%, and won two awards.

About

Hi, I'm Daniel.
Nice to meet you.

11 years across an airport, media, fintech, consumer apps, and now infrastructure. I work from data model to UI, use AI where it earns its place, and start by checking the real problem: the tickets, the people affected, and how we'll know the fix worked.

Side project

Full-stack product work outside my day job

California's housing crisis won't bend without more homes where people already live—and homeowners who can actually get an ADU built. adus.fyi exists to make that possible: cut through dense rules and guesswork, equip people to plan and permit accessory dwelling units with confidence, and turn backyard potential into real housing one project at a time.

  • I own architecture through release: implementation, testing discipline, App Store and web deploys, and post-launch fixes.

  • Two production surfaces—responsive web and a native iOS app—so the same product works on the desk and in the field.

  • Complex, jurisdiction-specific rules that push for clear models, validation, and maintainable feature growth—not a toy CRUD demo.

Tools I use most

TypeScriptNode.jsReactReact NativePythonTemporalPostgreSQLClickHousePrismaAWSLLM tool callingAI agentsMCPLLM evalsOpenTelemetryPlaywright