LP
Logan Park

Resume

Logan Park

Software Engineer | ML Engineer | US Citizen

Summary

ML-focused software engineer with 3.5 years of professional experience building production embedded systems and data pipelines in automotive telematics. Master's in Computer Science (ML/AI) from Arizona State University. Bachelor's in Computer Science with a Minor in Philosophy from the University of Colorado Boulder. I build end-to-end systems — from data ingestion and model development through deployment and UI. Interested in roles where I can work on data-driven decision tools, ML infrastructure, or applied AI.

Professional Experience

Software Engineer — Telematics Integration

LG Electronics Vehicle Solutions | Arizona | June 2022 — Present

  • Serve as the primary technical interface between OEM automotive manufacturers and internal engineering teams for embedded telematics devices deployed across North American vehicle fleets
  • Lead cross-company integration efforts to onboard OEM enterprise software onto proprietary telematics hardware, ensuring seamless operation across firmware, board support packages, and application software layers
  • Diagnose and resolve production issues across the full embedded stack — from cellular and BLE connectivity to application-level software — through on-vehicle triage, log analysis, and root cause investigation
  • Participate in architecture review meetings to design mitigation strategies for systemic device issues, collaborating across firmware, hardware, and OEM engineering teams
  • Conduct software cadence retrospectives analyzing feature implementation outcomes across production board releases
  • Validate software-hardware integration in the manufacturing pipeline (SMT, factory systems) before production runs
  • Train and support OEM engineers on device provisioning, software deployment procedures, and diagnostic workflows
  • Work directly with production vehicles — performing triage, debug, patch analysis, and issue reproduction in real-world conditions

Projects

LoL Draft Helper — ML-Powered Team Composition Analyzer

Python, FastAPI, Next.js, React, Recharts, Riot API, Pydantic Read the design article | Read Part 2

A team composition analysis tool for League of Legends solo/duo queue that profiles drafts across multiple dimensions and suggests picks backed by real match data.

  • Built end-to-end data pipeline ingesting 40,000+ ranked matches from Riot Games API with rate-limited continuous collection across 18 elo tiers (Silver through Diamond)
  • Engineered per-champion, per-position scoring system using real match performance data — CC output from timeCCingOthers, damage profiles from actual damage stats, tankiness from damage mitigated/taken, utility from healing output
  • Developed statistical profiling engine that computes team composition gaps against winning team profiles segmented by rank and patch, using mean and standard deviation for normalized gap analysis
  • Implemented three-list recommendation system: Best at Role (win rate + pick rate weighted), Best for Team Comp (gap-closing analysis), and Off-Meta (joker picks flagged by sample size threshold)
  • Built damage diversity scoring that boosts champions with true damage, %HP damage, or mixed damage profiles when teams are AD/AP imbalanced
  • Designed position-aware filtering using actual teamPosition match data instead of unreliable static role tags
  • Created interactive frontend with dual radar charts (team vs enemy), gap summary labels (Strong/Sufficient/Lacking), autocomplete champion search, ban tracking, and user role/duo inputs
  • Iterated through multiple versions based on real gameplay testing, identifying and fixing systematic issues: off-meta picks ranking too high, role-specific dimension weighting being too blunt, niche picks having inflated win rates due to specialist bias

GMS Scouter — MapleStory Boss Clear Calculator

TypeScript, Next.js, React, Recharts

A companion tool for MapleStory Global that calculates boss clearability based on your character's stats, class, and gear — filling a gap left by the absence of a GMS API.

  • Implemented MapleStory's full multiplicative damage formula: weapon constant, stat multiplier, attack power, damage/boss damage buckets, final damage, critical damage, skill damage%, IED vs PDR defense calculation
  • Built per-class mechanics data for all 40+ classes covering: main bossing skill damage%, clone/shadow multipliers (Shadow Partner 1.7x, Mir dragon 1.5x, etc.), boss PDR debuffs (Paladin Noble Demand 50%, Night Lord Frailty Curse 30%), conditional hyper passives, proc DPS contributions, and hidden Final Damage sources
  • Modeled HEXA Enhancement Cores (4 skills per class, universal FD scaling to Lv.30), V Matrix Boost Nodes (Lv.0-60, +2% FD/level), and HEXA Stat Nodes (3 configurable nodes with primary + secondary stats)
  • Built toggleable buff system covering all GMS consumables, guild buffs, and class buffs with per-buff stat contribution breakdowns
  • Implemented boss data for 40+ encounters across all difficulties (Chaos Zakum through Hard Baldrix) with verified HP, PDR, phase breakdowns, time limits, force requirements, party sizes, and crystal values
  • Added stat equivalence calculator showing marginal value of each stat type (ATT, %stat, %boss, %crit damage, %FD) relative to current build
  • Calibrated and validated against real endgame player data — calculator matches actual boss clear experience across 25+ encounters
  • Character profiles save to localStorage for persistent gear tracking and upgrade comparison

Automated Warehouse Planning System

Answer Set Programming, Clingo, Python | ASU Master's Project

  • Built logic-based multi-robot coordination system for warehouse grid navigation using declarative constraint programming
  • Designed mathematical termination and concurrency constraints to solve deadlock problems in simultaneous robot movement
  • System generated collision-free multi-robot path plans for all test scenarios across varying grid sizes and robot configurations

SmartHome Gesture Recognition Pipeline

Kotlin, Jetpack Compose, CameraX, Python, Flask, Docker | ASU Master's Project

  • Built end-to-end mobile-to-server data capture pipeline for gesture video collection, storage, and ML model training
  • Implemented Android camera pipeline with Jetpack Compose UI and HTTP upload to Dockerized Flask backend
  • Validated pipeline with pre-trained neural network gesture classification across 12 gesture categories

Blockchain Engineering — Public & Permissioned Chains

Solidity, Ethereum, Hyperledger Fabric, Go | ASU Master's Project

  • Implemented ERC-721 smart contract on Ethereum testnet exploring token ownership, transfer mechanisms, and on-chain state enforcement
  • Built Hyperledger Fabric permissioned network with Go chaincode for supply chain asset lifecycle tracking
  • Compared architectural tradeoffs between public (trustless, decentralized) and permissioned (governed, enterprise) blockchain systems

SortaLogic — Personal Portfolio & Content Platform

Next.js, TypeScript, Tailwind CSS, Vercel, Markdown

  • Built and maintain a personal platform for technical writing, project showcases, and product pages
  • Designed markdown-based content management system with YAML frontmatter, build-time parsing, and automated deployment via GitHub to Vercel
  • Published technical articles on ML system design, data engineering tradeoffs, and philosophy of technology

Education

Master of Science in Computer Science — ML/AI Focus Arizona State University — 2024

Bachelor of Arts in Computer Science — Minor in Philosophy University of Colorado Boulder — 2021

Technical Skills

Languages Python, TypeScript, JavaScript, Kotlin, Go, Solidity, SQL, C

Machine Learning & AI PyTorch, TensorFlow, Keras, scikit-learn, NumPy, pandas, feature engineering, model evaluation, classification, clustering, CNN architecture, statistical profiling, data normalization, min-max scaling, per-position scoring, gap analysis, recommendation engines

Data Engineering REST API data pipelines, real-time data ingestion, rate-limited collection systems, ETL, JSON/file-based data stores, batch processing, continuous collection scripts, data aggregation by segmentation (elo/patch/position)

Web Development Next.js, React, Tailwind CSS, FastAPI, Flask, Recharts, Pydantic, responsive UI design, dark theme design systems, autocomplete search, radar chart visualization

Mobile Development Android (Kotlin, Jetpack Compose, CameraX), mobile-to-server data pipelines, Docker containerization

Embedded Systems & IoT Telematics device integration, BLE/Bluetooth, cellular connectivity, firmware diagnostics, OBD-II protocols, CAN bus, BSW/application layer debugging, on-vehicle triage

Blockchain Ethereum (Solidity, ERC-721), Hyperledger Fabric (Go chaincode), smart contract development, permissioned vs public chain architecture

Tools & Infrastructure Git, GitHub, Vercel, Linux, Docker, Clingo/ASP, Hyperledger Fabric, httpx, uvicorn

Concepts Multiplicative damage formulas, equivalent stat conversion, decision matrix design, team composition profiling, position-aware winning profiles, joker pick detection, archetype analysis, boss clear calculation, per-class mechanics modeling

Languages

  • English — Native
  • Korean — Bilingual
  • Japanese — Currently learning

Writing & Publications

What I'm Looking For

ML Engineer, Software Engineer, or AI Engineer roles where I can build data-driven systems that help people make better decisions. Interested in teams working on applied ML, data infrastructure, recommendation systems, real-time analytics, or human-centered AI tools. Open to work in all locations.

For a condensed one-page version or to discuss opportunities, reach out via the contact information on the homepage.