Skip to content

Resume

michael@mta.io | values | blog | linkedin | github | gists

kind, thoughtful, and bright team player obsessed with growth, user satisfaction, and getting the job done

Knowledge

Expert Advanced Proficient
Languages JavaScript/TypeScript & Python Ruby & PSQL/SQL Java/Kotlin & NoSQL
Frameworks NodeJS Electron, Express, Flask, & Rails Prisma
Libraries Jest RSpec & Vue React, Pandas, & NumPy
Paradigms SOLID, OOP, & TDD Agile, CI/CD, On-Call, & REST CRUD
Tech Firebase AppEngine Docker, MongoDB, AWS S3, & AWS Lambda

Impact Summary

  • Medallion, Senior Software Engineer (1 year) - HealthTech

    • Built automations and integrations to third party websites (NPPES and CAQH) that reduced or eliminated manual data validation and insertion.
    • Leveled up two junior engineers e.g. implementing a finite state machine for data validation and adopting frontend testing.

  • OmniMerchant, Co-Founder (2 years) - FinTech Startup

    • Created a desktop app that reduced the time to process payments by 75% and reduced transaction fees by 30%.
    • Managed a 5-member team of engineers, contractors, and designers to ship designs, integrations, and features.

  • Stitch Fix, Data Platform Engineer (1 year) - Online Retailer, details

    • Implemented and launched a successful experiment that increased revenue by $750K monthly.
    • Refactored the core recommendation engines that received millions of requests to be more resilient and transparent.

  • Stitch Fix, Senior Software Engineer (3 years) - Online Retailer

    • Researched, designed, and led new payment terms for purchase orders - saved $30M annually.
    • Launched the UK vendor data platform and added a client feedback stream - attracted more vendors and prevented a lawsuit.

  • Ordino, Co-founder (2 years) - AgTech AI Startup, learnings

    • Created an Android app that collected crop data from in-field sensors using drones - covered 90 acres in minutes vs hours.
    • Created Random Forest models to predict if a field will have above or below-average yield - 82% accuracy (F1-Score).

  • Enfant, Co-founder (1 year) - Infant Care AI Startup, always on-call, demo and summary

    • Developed Android and Google Home apps to translate baby cries - over 100 downloads and many relieved parents.
    • Created a system to collect and label training data in production - significantly increased training data size.

Experience

(full list on linkedin)

Medallion (Senior Software Engineer, 6/24 – Now)

bi-monthly week-long on-call rotation Languages: Python & TypeScript Libraries: Django, Protobuf, Pydantic, & React
Tools: DataDog, Docker, MetaBAse, PSQL, & Sentry

  • Designed and built an integration to NPPES (government API for healthcare providers) that automatically pulled data from the API, compared it to internal data, and informed the provider if the data was mismatched - this was a manual task before, now only needs human intervention 96% of the time.
  • Designed and built the Protobuf and Pydantic models and transformations for automating CAQH (source of truth for healthcare provider data) filling - 91% success rate on jobs, significantly reducing human intervention.
  • Designed and built the metrics, logs, and dashboards for measuring success and debugging the CAQH automation - wrote an internal blog on how to best use metrics/logs to correlate data in DataDog and how to best store Protobuf in PSQL for analysis.

OmniMerchant (Co-Founder, 8/22 – 5/24)

Languages: C++ & JavaScript/TypeScript
Libraries: Electron, Express, NodeJS, & Vue
Tools: AWS API Gateway, AWS Lambda, AWS S3, MongoDB, & Retool

  • Led weekly meetings with stakeholders, interviewed potential customers, created roadmaps, & hired then managed contractors.
  • Created clean abstractions for integrating with 3 payment gateways using 1 interface and set of interface tests - reduced integration time from 6 months to 4 weeks.
  • Built and deployed an LLM chatbot using Retrieval-Augmented Generation (RAG) LangChain client to answer business questions.
  • Created a monorepo to host and share multiple utility packages, the REST API, business logic, and website.
  • Created a Native NodeJS package that used the Node API to convert synchronous requests to asynchronous requests - made the app 100% faster by removing load screens.
  • Created a self-service Admin panel for customer and product key management - reduced installation and expansion costs.

Stitch Fix (Data Platform Engineer, 8/21 – 8/22)

bi-monthly week-long on-call rotation
Languages: Python & SQL
Libraries: Airflow, Docker, FastAPI, Flask, Numpy, Pandas, & Spark
Tools: AWS EC2, AWS S3, & DataDog

  • Worked with data scientists and business stakeholders to launch multiple experiments.
  • Created a config-driven recommendation platform that reduced maintenance and increased transparency for the core recommendation engines that serve every customer - merged 3 apps into 1 and enabled self-service debugging, details.
  • Migrated 3 critical recommendation engines from batch to real-time features - unlocked new features and removed deprecated dependencies.

Stitch Fix (Senior Software Engineer, 6/18 – 8/21)

monthly week-long on-call rotation
Languages: JavaScript, PSQL, & Ruby
Libraries: gRPC, Kafka, RabbitMQ, Rails, & React
Tools: AWS S3, BugSnag, DataDog, Docker, ElastiCache, & Firebase

  • Maintained the purchase order service, vendor service, and vendor sales dashboard, critical for purchasing, sourcing, and compliance.
  • Created a self-service system for the quarterly and yearly vendor manual and Terms & Conditions updates - the compliance teams can now do their jobs without needing engineering.
  • Implemented data-driven purchasing through a message queue and nightly jobs - reduced purchasing and inventory costs.
  • Implemented a feature that captures the country of origin for each purchase - improved tariff cost estimates and item sourcing.
  • Helped integrate the purchase order system with Kafka pub/sub topics - maintained SOX compliance.
  • Launched EDI integration with SPS and Oracle; led the bug resolution and purchase order conversion teams - prevented an audit, unlocked more buying options, and lowered the cost of each purchase order.
  • Created gRPC endpoints for getting and updating purchase orders - improved SOX compliance.
  • Created dozens of pages of documentation explaining the purchase order lifecycle in Stitch Fix's systems and how to fix broken POs.

Ordino (Co-Founder, 8/17 – 5/19)

Languages: Java, JavaScript, Kotlin, & Python
Libraries: Flask, Numpy, Pandas, & SciKit
Tools: Android, AppEngine (GPC), Firebase, & Google DataStore

  • Met with local tomato farmers to understand and design a solution to give them more insight into their crop quality.
  • Designed and created a dashboard to view weekly sensor readings and yield predictions.
  • Designed and created a system to read soil sensor data from a drone and tag it with the geolocation of the drone.

Enfant (Co-Founder, 4/17 – 6/18)

Languages: Java, Kotlin, & Python
Libraries: Flask, Numpy, Pandas, & TensorFlow
Tools: Android, AppEngine (GPC), Firebase, & Google DataStore

  • Met with new parents and babysitters to figure out how to best deploy the baby translator model and develop new features.
  • Created a system that allowed users to rate translations and tag training data with that rating.
  • Optimized and deployed a TensorFlow model on an Android app - reduced model size by 90% with minimal accuracy reduction.

Projects

HealthyHost (Lead Android Developer & Teacher, 8/17 – 5/18)

Healthcare App (Java)

Taught and mentored a group of first-time programmers on Android development, version control, and GitFlow.
Implemented the media playback feature to play audio translations of common medical ailments in 4 different languages.

Writing

  • Maggie: Building A Baby Translator - blog
  • Python: 30 Questions to Learn Multithreading - blog
  • Migrating Stitch Fix's Core Recommenders to a Config-Driven Platform - blog
  • I Built Our Admin Panel in Retool - Being Agile with Lowcode - blog
  • PSQL: How to Use HAVING to Query Aggregate Functions and Groups - blog
  • Python: Why You Can't Use Arrays in Query Strings with Requests and urllib - blog

Volunteerism

  • Sideshow Abatement/Traffic Calming - Organizing Neighborhood (2024 - Current) - lobbied the city of Oakland to pass $1.4 million amendment for sideshow abatement and traffi calming
  • Trees for Oakland - Tree Planting Lead (2018 - Current) - planted hundreds of trees and organized tree plantings
  • Friends of the Urban Forest - Volunteer & Tree Planting Lead (2019 - Current) - planted dozens of trees and completed urban forester class

Education

Bachelor of Science (BS), Computer Science and Engineering - UC Merced, 2018