Michael DzobaData engineer who builds systems people actually use

Eight years turning messy data into pipelines, dashboards, and decisions, most recently at GE Aerospace and Progressive Insurance. I also built Quanta Analytics, an analytics SaaS for game servers that's now in closed beta. I'm looking for my next data engineering or analytics role.

$13M+in potential duty savings surfaced by dashboards I built at GE Aerospace
40+ hrsof manual work automated away every month across two employers
1.8Mgame events logged by Quanta Analytics, the SaaS I built
4.94★across 224 guest reviews as a short-term rental host

Selected work

Projects I designed, built, and still run. The code for these is private, so each one is shown through what it does, how it's put together, and what it produced.

Quanta Analytics

Free closed beta since October 2026

A multi-tenant analytics SaaS for Rust game servers. It captures everything that happens in-game and turns it into 20+ dashboard pages on player behavior, retention, cheating, and map balance.

1.8Mevents logged
369players tracked
3,545hours of play recorded
5servers connected

The problem

Rust server owners make big calls about loot, events, map design, and moderation with almost no data. The game fires events constantly, but nothing collects, stores, or explains them over time.

What I built

  • The entire data path, from the in-game plugin to the ingestion API, database design, and dashboard.
  • Session-based event tracking: the API ties every event to the player's play session and the server's current wipe, so each session can be replayed and wipes compared side by side.
  • Time-series storage on TimescaleDB: event tables split into daily chunks that compress after 24 hours, with hourly continuous aggregates behind the heavier charts.
  • Query tuning so each page reads each table only once, plus rollups and response caching, benchmarked against a 1.5M-row copy of production.
  • An opt-in public leaderboard API that exposes only a random public ID, which now powers the leaderboards on thelastofrust.com.
  • Multi-tenant accounts: Steam and Discord sign-in, organizations that own several servers, invite links, and roles that control which pages each team member sees.
  • Separate production and development environments, each with its own branch, service, and database.

What server owners can see

  • HeatmapsPvP, building, looting, and gathering drawn over the real game map
  • Player sessionsEvery kill, loot, craft, and chat line on a timeline
  • RetentionDay 1, 3, and 7 retention and each wipe's half-life, compared wipe over wipe
  • Cheat detectionAccuracy and headshot outliers by weapon, alt accounts, and Steam ban checks
  • Map balanceRisk versus reward for every monument, and NPC fight outcomes
  • Staff and moderationAdmin logs, staff coverage by hour, and report response times

Stack

  • C#
  • Python
  • FastAPI
  • asyncpg
  • PostgreSQL
  • TimescaleDB
  • Jinja2
  • Chart.js
  • Leaflet.js
  • nginx
  • Ubuntu
  • systemd

quantaanalytics.gg

The Last of Rust

Game servers and community site

Three heavily modded Rust servers, two PvE and one PvP, that I run end to end: hardware, custom gameplay code, the website, and the community of 800+ Discord members.

Infrastructure

I ran a colocation-versus-rental cost analysis and moved the servers onto two dedicated machines I picked for Rust's heavy CPU load: AMD Ryzen 9800X3D processors with 64 GB of RAM each.

Favorite plugins I've built in C#

Object-oriented C# built to be configured rather than edited: each plugin reads JSON config files, so server owners can tune its behavior without touching the code.

  • Drug Deals: lets players buy, grow, and sell drugs for in-game currency. Seeds arrive by cargo plane, and harvests vary with each plant's genetics.
  • Drug Raids: lets players raid drug hideouts defended by turrets and NPCs, stealing drugs they can then sell through Drug Deals.
  • Wonder Weapon: a ray gun in the style of Call of Duty Zombies. It repurposes a chosen weapon so it can gather resources and kill zombies.
  • Server Menu: a custom in-game UI that puts the server's rules, features, leaderboards, and helpful info in one place for players.
  • Player Spawn Points: sends players to different spawn locations based on their permissions and whether they're new or returning. Monument spawns are stored relative to the monument, so they survive map wipes.

Drug Deals and Drug Raids both build on a shared Drugs API plugin I wrote, so the drug types and plant genetics are defined once and reused.

I also sell plugins on Codefling, the main marketplace for Rust plugins, where 150+ copies of my 5 paid plugins have sold to other server owners.

Stack

  • C#
  • Object-oriented design
  • JSON configs
  • Oxide / uMod
  • Unity
  • Dedicated hardware
  • Astro
  • TypeScript
  • Cloudflare Workers
  • Steam Web API

The website: thelastofrust.com

A static Astro site on Cloudflare with server guides, rules, a feedback form, and store integration. A small Worker handles only the live data, and its leaderboards are fed by Quanta Analytics.

  • Secure by design: server owners switch the public feed on from the dashboard, which generates a random public ID, so internal server IDs and API keys are never exposed.
  • Fails gracefully: if one server's feed is down, that server comes back empty and the others still show. If every feed fails, nothing is cached, so the next visit tries again.
  • Live player counts from Steam's public server list, cached for 60 seconds.
  • Same-origin API, so the browser needs no cross-origin setup, and images hosted on Cloudflare work both on the site and in-game.

Short-term rental business

3.5 years hosting

I bought and operate a short-term rental, from the investment analysis to the day-to-day guest experience.

  • Ran market and financial analysis to choose the property and project its returns.
  • Adjust pricing across seasons and local events to keep occupancy high.
  • Handle all operations: guest communication, turnovers, and maintenance.

4.94 rating across 224 guest reviews.

Experience

Data work across aerospace, insurance, and hockey.

  1. Data Engineer

    Oct 2022 to Jun 2025

    General Electric Aerospace

    • Built Spotfire dashboards that surfaced $13M+ in potential import/export duty savings.
    • Built production data pipelines in AWS Redshift, cutting Spotfire report load times by about 20%.
    • Automated ingestion of structured and unstructured data with SQL, Python, Starfish, and Dataiku, eliminating 10+ hours a month of manual uploads.
    • Wrote automated SQL and Python tests covering tables, views, functions, and connections.
    • Designed cross-database reporting pipelines for access control and compliance, and delivered audit-ready reports on document access.
  2. Intern to Data Analyst III

    May 2018 to Oct 2022

    Progressive Insurance

    1. Data Analyst IIIClaims Control, Medical teamNov 2021 to Oct 2022
    2. Data Analyst IIAcquisition, CRM Business Intelligence, IT ControlFeb 2020 to Nov 2021
    3. Data Analyst ICenter of ExcellenceFeb 2019 to Jan 2020
    4. InternData Management & AnalyticsMay 2018 to Aug 2018
    • Automated monthly reporting workflows in Alteryx with end-to-end ingestion, logging, and transformation pipelines, eliminating 30+ hours a month of manual work.
    • Applied NLP and n-gram analysis in Python to measure customer demand for annual policies, uncovering interest across auto, pet, and life insurance.
    • Analyzed customer browser and OS combinations with SQL and Python to prioritize support for the configurations that mattered most and reduce quote drop-offs.
    • Built SSIS ETL pipelines feeding AWS-hosted data science models for real-time policy scoring.
    • Evaluated 30,000+ technical documents with C# and SQL, informing a decision to avoid an unnecessary software purchase.
    Show all details

    Data Analyst III

    • Delivered personal injury protection (PIP) analytics for Michigan, plus ad hoc analysis for the Claims Control Medical team.

    Data Analyst II

    • Automated monthly reporting workflows in Alteryx with end-to-end ingestion, logging, and transformation.
    • Built and maintained SSIS ETL pipelines feeding AWS-hosted data science models for real-time policy scoring.
    • Evaluated 30,000+ technical documents with C# and SQL to measure content reuse.
    • Used NLP and n-gram analysis in Python to measure demand for annual policies, uncovering interest across auto, pet, and life insurance.
    • Ran A/B tests and funnel analysis in SAS and Excel to optimize the quote flow, identifying the most effective placement for the Help button.
    • Analyzed customer browser and OS combinations with SQL and Python to prioritize support for the configurations behind quote drop-offs.
    • Overhauled the enterprise timesheet compliance report and automated refreshes for the LEAN dashboard with SQL, Alteryx, and Tableau.

    Data Analyst I

    • Led the transition from Excel to Tableau for enterprise-wide reporting, improving visibility across business units.
    • Completed the Google Cloud Platform Specialization.

    Intern

    • Built predictive models with TensorFlow neural networks in GCP BigQuery to forecast customer behavior and insurance trends.
    • Built an internal chatbot with Python, Flask, and NLTK so employees could self-serve data requests.
  3. USA Hockey Coach

    Jan 2019 to Mar 2022

    Team Ohio High School

    • Built a Python application that stored the team's data in a SQL Server database: schedule, player points, faceoff wins and losses, plus/minus, and shot locations.
    • Ran practice drills focused on developing every part of the forwards' game.
  4. Assistant to the Director of Hockey Operations

    Aug 2015 to Dec 2018

    Miami University Division I Hockey

    • Recorded live game data and tagged game and practice footage of Miami and its opponents, which the coaching staff used for teaching and scouting.

Skills, and where I've used them

Every tool here has shipped something real, at work or in one of the projects above.

Data engineering

SQL, AWS Redshift, S3
Production pipelines at GE Aerospace
PostgreSQL, TimescaleDB
Quanta's time-series storage, compression, and hourly rollups
MS SQL Server, SSIS, Oracle
ETL into real-time scoring models at Progressive
Dataiku, Alteryx, Starfish
Automated ingestion and reporting, 40+ hours a month saved

Analytics and visualization

Spotfire
Dashboards behind $13M+ in identified savings
Tableau, Looker, Power BI
Led Progressive's move to Tableau
Leaflet.js, Chart.js
Heatmaps and trends in Quanta Analytics
GCP BigQuery, TensorFlow
Customer behavior forecasting

Software development

C#, JSON
Object-oriented game server plugins with JSON configs, 150+ copies sold on Codefling; Quanta's event logging; document analysis at Progressive
Python, FastAPI, Flask
Quanta's async ingestion API; internal data chatbot at Progressive
JavaScript, TypeScript, Jinja2
Quanta's 20+ dashboard pages; the TLOR leaderboard Worker
Git, GitHub
Version control across all of the above

Infrastructure and operations

Linux, nginx, systemd
Running Quanta's production and development environments
Dedicated hardware
Sizing, cost analysis, and CPU tuning for the game servers
Astro, Cloudflare Workers
thelastofrust.com and its live leaderboard API
Business operations
Running a 4.94-rated rental, from pricing to guest relations

About me

I like problems where the data is messy and the answer changes what someone does next. That's true whether it's import duties at GE, quote drop-offs at Progressive, or where players spend their time on a game map.

I grew up partly in Japan, living there for ten years and going to a Japanese junior high school, which taught me to notice how I come across to others and adjust, a habit that's served me well ever since. I've played ice hockey for most of my life, coached high school hockey for three years, and now play broomball in several leagues in Minneapolis.

Education

MS, Software DevelopmentBoston University, 2022

BA, Mathematics, minor in Computer ScienceMiami University, 2018

Let's talk.

I'm available now for data engineering and analytics roles. The fastest way to reach me is email.