Dustin Stewart
Software engineer building production systems, zero to one.
8+ years shipping 0→1 across AWS, Azure, and GCP — APIs, customer-facing apps, AI platforms, and the CI/CD behind them.
From API design to the infrastructure underneath.
I'm a senior engineer who builds production systems from 0 to 1 — equally at home designing APIs, shipping customer-facing apps, and standing up the CI/CD and cloud infrastructure that carries them.
Lately that's founding-level fullstack and AI-platform work for early-stage startups, after leading fullstack engineering for the LA Galaxy and Fortune-500 clients at 55. My roots are in data and ML, and it still shows in how I think about systems.
- Based in
- Los Angeles, CA
- Experience
- 8+ years · 0→1 across AWS, Azure & GCP
- Focus
- Fullstack · AI platforms · data & cloud
- Certified
- AWS Pro Solutions Architect · GCP Pro Data Engineer
- Education
- USC — B.S. Data Analytics
What I'm building lately.
Outside the day job I'm usually elbow-deep in a homelab or wiring an AI agent into something it has no business touching yet. I like owning the whole stack — from the silicon in my closet to the pixels on the screen.
Updated June 2026- Leading backend & AI-platform engineering at Lightward.
- Running a multi-service homelab on a refurbished ThinkCentre — home automation, cameras, backups, and this very site — all Dockerized behind Tailscale & Cloudflare.
- Building agentic tooling with Claude + MCP for personal automation.
- Ex-music-industry — still have loud opinions about the Gorillaz catalog.
Where I've worked.
Lead Software Engineer · Lightward
Architected and built the production backend platform on AWS — FastAPI on ECS, async Lambda, DynamoDB, OAuth, and EventBridge — with KMS envelope encryption, RBAC, audit logging, and full Terraform IaC. Built a centralized AI gateway for Claude / OpenAI / Llama (audit trail, tiered authorization, rate limiting, prompt-injection defense), a consent-driven permission model, and an in-house observability + threat-detection stack.
Senior Fullstack Engineer · The Los Angeles Galaxy
Built the club's first in-house Transfer Analytics Platform (React + FastAPI + Databricks + Redis) serving real-time scouting intelligence to the GM and coaching staff. As a founding AEG Sports engineer: stood up two Databricks medallion warehouses, DLT/PySpark pipelines, and Unity Catalog; shipped 5+ internal apps, an XGBoost player-valuation model, and AEG's first CI/CD & DevOps practices.
Lead Engineer · Senior ML Engineer · 55: The Data Company
Led architecture and delivery of data platforms across AWS/GCP (Snowflake, BigQuery, Redshift) — work that drove an 89% increase in cloud-services revenue. Built end-to-end ELT (Databricks, Kafka, Airflow), production ML serving on Vertex AI / Kubeflow, and a multi-model LLM experimentation app, while owning containerized infra, IaC, and CI/CD across US clients.
ML & Analytics Consultant · Self-Employed
Built analytics pipelines, marketing-attribution models, and reporting automation in Python and SQL; advised clients on ML prioritization and feature engineering to maximize KPIs.
Engineering Intern · Warner Music Group
Frontend development, data collection, and API integrations for large-scale digital campaigns and assets for global artists including Dua Lipa and Gorillaz.
Recent projects.
AI gateway @ Lightward
A centralized gateway for public + internal LLM usage (Claude, OpenAI, Llama) with a full audit trail, tier-based action authorization, per-user rate limiting, prompt-injection blocking, and a live performance + security dashboard.
Vidpod
A video-podcast studio: an episode designer with a clip + audio-waveform timeline, ad A/B testing, and an analytics dashboard tracking impressions, CTR, and revenue across published episodes.
spotiflair
Turns a Spotify playlist into a pretty, themeable, screenshot-ready view for sharing. Static front end on CloudFront with a serverless API.
wc-share
Paste a FIFA World Cup bracket link and it renders a clean, themeable, screenshot-ready bracket image to share. A Fastify backend pulls the bracket JSON; the front end themes it and exports a PNG.
Self-hosted home lab
A multi-host home lab on Tailscale: Dockerized services behind tunnels, infrastructure-as-code, with health monitoring and encrypted S3 backups. Where this very site is hosted.
What I work with.
Backend & APIs
Frontend
Cloud & Infrastructure
Data Engineering
AI / ML
Let's talk.
Have a problem worth solving, a role, or just want to compare homelab notes? My inbox is open.
[email protected]