SYS.COORD LAT 37.7749 / LNG -122.4194
SECTION
SCROLL.DEPTH 0.00%
TIMESTAMP 00:00:00

YONGJOON

.dev

researcher / developer / architect of systems

v3.2.1 EST. 2019 SEOUL / SF

RESEARCH LOG

[01] SYSTEMS

Distributed Computing

Exploring fault-tolerant architectures for real-time data processing. Building resilient systems that adapt and self-heal across heterogeneous infrastructure. Focused on consensus algorithms and event-driven microservice patterns that scale from edge to cloud.

2024—PRESENT ACTIVE
[02] ML/AI

Neural Architecture Search

Investigating automated methods for discovering optimal network topologies. Combining evolutionary strategies with gradient-based optimization to find efficient architectures for resource-constrained deployment scenarios.

2023—PRESENT ACTIVE
[03] LANGUAGE

Compiler Design

Designing domain-specific languages and intermediate representations for heterogeneous compute. Exploring type-theoretic foundations for memory-safe concurrent programming with zero-cost abstractions.

2022—2024 ARCHIVED

PROJECT ARCHIVE

Nexus Runtime

High-performance distributed task orchestration engine with adaptive load balancing

Rust / gRPC / WASM

Prism Compiler

Multi-target DSL compiler with algebraic effect system and linear type inference

OCaml / LLVM / SMT

Orbit Graph DB

Temporal knowledge graph with causal inference engine and real-time query planning

Go / RocksDB / GraphQL

Signal Monitor

Observability platform with anomaly detection using learned baseline behavior models

Python / ClickHouse / React

Wavelet NAS

Neural architecture search framework using multi-objective evolutionary optimization

PyTorch / Ray / ONNX
1 / 5

TERMINAL ARCHIVE

Publications & Writings

  • 2025.11 On Convergence Rates of Federated Averaging Under Non-IID Data Distributions ICML Workshop
  • 2025.06 Efficient Sparse Attention Mechanisms for Long-Context Transformer Architectures NeurIPS
  • 2024.09 Type-Safe Distributed Protocols via Session Types in Practice OOPSLA
  • 2024.03 Memory-Efficient Graph Neural Networks for Billion-Scale Knowledge Bases VLDB
  • 2023.12 Compositional Program Synthesis from Natural Language Specifications PLDI
  • 2023.07 Causal Inference in Temporal Knowledge Graphs: A Structural Approach KDD
“The most interesting problems in computing are those that sit at the boundary between the formally provable and the empirically observable.”

About

Yongjoon is a systems researcher and software engineer focused on the intersection of programming languages, distributed systems, and machine learning infrastructure. With a background spanning compiler design and large-scale data processing, the work centers on building tools that make complex systems more understandable, reliable, and efficient.

Currently exploring the design space of domain-specific languages for heterogeneous compute environments, and investigating how formal verification techniques can be made practical for everyday distributed systems development.

EDUCATION Ph.D. Computer Science
FOCUS PL / Systems / ML Infra
LANGUAGES Rust, OCaml, Python, Go

Building systems that think,
in languages that prove.

© 2026 Yongjoon All systems nominal