ENGINEER
Software Engineer · San Francisco, CA

Andy Tran

Backend engineer. AI tinkerer. I build the systems that run quietly in the background and hold up when it matters.

Andy Tran
Scroll

I build things
people actually use.

SF native. Software engineer who's spent the last few years building things that actually run in production: ad infrastructure for live TV at Red Bull Media House, enterprise pricing software at Supplyframe (now part of Siemens). Real problems, real scale, and yeah, things broke sometimes.

Lately I'm deep into AI tooling. Automating real workflows with Claude and Whisper, not just shipping demos. I like backend work that has to hold: data pipelines, streaming infra, systems where getting it wrong means a 3am page.

When I'm away from a keyboard I'm volunteering with MissionBit, teaching intro coding to SF high schoolers since 2019. Turns out explaining Git to a 16-year-old is harder than building a microservice. My 14-year-old dog Maddy keeps me grounded through all of it.

0
Records processed daily
0
Device types shipped
0
Students mentored
0
Countries expanded to
TypeScript JavaScript Python SQL Ruby Node.js Express React Vue.js PostgreSQL MongoDB BigQuery Redis Firebase Prisma GCP AWS S3 Docker CI/CD Claude API OpenAI API Whisper RAG Pipelines Agentic Systems REST APIs Elasticsearch HLS/DASH Jest Zod TypeScript JavaScript Python SQL Ruby Node.js Express React Vue.js PostgreSQL MongoDB BigQuery Redis Firebase Prisma GCP AWS S3 Docker CI/CD Claude API OpenAI API Whisper RAG Pipelines Agentic Systems REST APIs Elasticsearch HLS/DASH Jest Zod

Where I've
left a dent.

Aug 2022 - Feb 2024
Supplyframe · Siemens
Software Engineer, Full Stack

Supplyframe built software for the electronics supply chain, acquired by Siemens while I was there. I worked on the CPQ product: the pricing and quoting engine that enterprise distributors used to generate quotes across millions of components. When pricing logic breaks, deals break. Learned to care a lot about correctness.

  • Built the core pricing engine powering quoting, calculations, and config for 3 enterprise distributor clients across 50K+ SKUs
  • Optimized PostgreSQL pipelines handling 200K+ records, cutting query execution time by 20%
  • Wrote integration tests that caught 30+ bugs pre-production and reduced QA cycle time by 15%
  • Automated enterprise reporting workflows, cutting generation time from 2 hours to 15 minutes
Oct 2020 - May 2022
Red Bull Media House
Software Engineer, Backend

ServusTV and Red Bull TV are live broadcasters streaming to millions of devices worldwide. I joined as the sole engineer responsible for the entire ad delivery layer: the system deciding what ad plays on what device, in what format, at what time, across 23 different platforms simultaneously. Live TV doesn't pause for bugs. Neither did I.

  • Sole engineer on the ad microservice, shipped across 23 device types and multiple third-party ad SDKs
  • Built async pipeline processing 5M+ device records/day with BigQuery analytics and StackDriver observability
  • Geo-streaming rollout to 2M+ devices across 3 new countries
  • Migrated internal CMS to Vue.js and GCP App Engine, cutting deploys from 45 minutes to under 15
  • Owned live and VOD playback (HLS/DASH) with custom ETL pipelines for 24/7 streaming delivery
2019 - Present
MissionBit
Volunteer Instructor & Curriculum Dev
  • Taught Git, Python, and web development to 100+ high school students with 1:1 mentorship on career paths into software engineering
  • Led JavaScript curriculum redesign during transition to online learning

Things I've built
outside of work.

02
AI · Personal Tool
AI Assistant Orchestrator
Personal AI assistant controlled via SMS. Routes natural language commands through Claude API with a tool allowlist and Zod validation to prevent unauthorized execution. Sub-agent delegation for specialized tasks with confirmation workflow for sensitive actions. Full audit trail via Prisma/SQLite.
Claude API Twilio Express Prisma Zod
03
Integration · Production
Discord Sheets Bot
Discord → Google Sheets, automated. 200+ users, runs in production, zero babysitting.
Node.js Discord.js Google Sheets API

Not everything
ships to prod.

Gaming
Competitive history

Took 2nd at the Soldier Front national tournament in 2009. One map away from first place and representing the United States at worlds in Thailand. Still think about it. From there: high school LoL team captain, IPL San Francisco qualifier, eventually Master tier. Twisted Fate one-trick. No regrets. Okay, maybe one.

Soldier Front
2nd Place, National Tournament · 2009
M
League of Legends
Master Tier S5 & S6 · top 0.01% High Diamond S3 & S4 · Top 500 Season 1 HS team captain · IPL SF qualifier Twisted Fate op
Teaching · Volunteering
MissionBit

Teaching intro coding to SF high schoolers since 2019, that's 100+ students across six years. Turns out explaining Git to a 16-year-old is harder than building a microservice, and I mean that as a compliment to both.

Python JavaScript Web Design Figma
Since 2019 · San Francisco · 100+ students Still the best thing I do outside of work.
Basketball
Lakers fan

Been a Lakers fan since I watched Kobe and Shaq as a kid and never looked back. Luka + LeBron is the most fun this team's been in years and I genuinely think this is a title window. Yes I've been burned before. Yes I'm still watching every game. Some loyalties aren't rational.

23 LEBRON 23 77 LUKA 77
Travel
Places I've been

Always looking for the next city. Domestic or international, doesn't matter. Every trip makes me a better engineer somehow, or at least that's what I tell myself.

🇦🇺 Melbourne 🇦🇺 Gold Coast 🇬🇧 London 🇻🇳 Hanoi 🇻🇳 Nha Trang 🇹🇭 Bangkok 🇲🇽 Cabo San Lucas 🇯🇲 Kingston 🌺 Honolulu, HI 🗽 New York City 🤠 Austin, TX 🌲 Seattle, WA 🏔 Anchorage, AK 🌊 San Diego, CA 🎰 Las Vegas, NV ⛷ Lake Tahoe, CA 🐻 Big Bear, CA 🎲 Reno, NV
Side Project · ML
Disney World Wait Times

I've been to Disney World once. Got curious about wait times, found a 7-year public dataset, and spent a weekend with a team of 4 building models that predict queue times by attraction, weather, and school calendar. The random forest outperformed linear regression by a lot, which was the most satisfying part.

Python pandas scikit-learn Matplotlib Random Forest Linear Regression
View on GitHub →

Where it
started.

2016 - 2019
B.S. Applied Statistics
University of California, Santa Barbara
Relevant coursework: Graduate-level Advanced Machine Learning
2019 - 2020
Full Stack Software Engineering
App Academy
1,000+ hour immersive program focused on web application development