Junghyun Nam (남정현)
Last updated: 2023-05-15
Also available on https://cro.sh/me
I'm living in Seoul, Korea.
For now I'm not actively searching jobs but inbound job offers are appreciated.
I'm on an alternative supplementary service as skilled industrial personnel, ending on July 30th, 2023. Hence I can only work full-time on designated companies non-remotely by the Office of Military Manpower Administration of Korea.
저는 현재 보충역 산업기능요원으로 복무 중이며 예상 소집해제일은 2023년 7월 30일입니다.
Seoul National University 2019.03 ~ leave in absense
Currently on undergraduate courses majoring in Computer Science & Engineering.
- GPA: 3.9/4.3 (not finalized)
- Credits: 45/41 (not finalized)
Also double-majoring in Statistics.
Sejong Science High School 2016.03 ~ 2019.02
Hyperithm (https://hyperithm.com) 2021.08 ~ current
Hyperithm is a leading digital assets manager in Asia, based on Japan and Korea.
- Head of Infrastructure
- Maintaining the overall cloud infrastructure based on AWS and GCP.
- Developed and maintaining a HFT-optimized cryptocurrency exchanges data feed.
- Keywords: Asynchronous Programming, Rust, Tokio, HTTP, WebSocket.
- Designed & managed development of a trading strategy management
- (think stripped-down version of ECS or EKS with some internal niche feature demands).
- Keywords: Rust, AWS CloudFormation, AWS S3
- Developed & maintaining an DevOps toolkit for AWS infrastructures, featuring seamless 1Password integration with AWS credentials management.
Research Intern at CARES Lab 2021.03.26 ~ 2021.06.30
Started as SNU UROP(undergraduate research opportunities program) on February, then switched to paid intern on March.
Worked on RustFTL, a (partial) Rust port of the FEMU SSD emulator.
Rewrote FTL(flash translation layer) part of FEMU into Rust and reported about use of Rust on embedded systems focusing on safety, performace impact, and idiomacy of Rust code on FTL structures.
The final report and documentation were constructed with mdBook and rustdoc.
Engineering Intern at CN.AI 2020.08.01 ~ 2020.08.28
Constructed data pipeline for deep audio classifier with PyO3 and HDF5 to boost training speed.
Developed a demo website with StyleGAN2 which synthesizes 'median face' from given two facial images.
- The model is pretrained with FFHQ dataset on TensorFlow, and communicates with backend with a simple TCP server.
- The backend is written in Golang with its standard
net/httpmodule and the Gorilla web toolkit.
- The frontend is based on React.
- Rust(primary): Tokio and related async ecosystem packages, aws-sdk-rust
net/http, Bolt(embedded KV store)
- TypeScript: Mostly React
- Python: Tkinter, requests, Bolt(Slack SDK)
- Other misc. languages w/ basic experiences: C++, Java(JVM, Android), Dart(pre-2, Flutter), OCaml
Have managed several on-premise solutions such as Jenkins, a security management platform (private)
- CI/CD pipeline with AWS EC2, EC2 Image Builder, ECR, etc.
- Layer-7 HTTP proxy clusters with AWS EC2 Launch Template, ECS, ECR, etc.
- Fine-grained access control of EC2 instances with AWS Systems Manager Session Manager and IAM.
Linux & Systems Programming
- Embedded Linux development with Raspberry Pi(1B+, Zero W, 3B+) and Buildroot.
- Kernel hacking including implementing custom syscalls and task schedulers.
Blockchain & DeFi
- General consensus mechanisms, e.g. Proof-of-Work, (naïve) Proof-of-Stake and their variants like Proof-of-History, Byzantine Fault Tolerance, etc.
- How decentralized finance works and their sub-categories: money markets, DEXes, yield aggregators, stablecoins, etc.
- Cross-chain/multi-chain/layer-2 ecosystems: Wormhole, Cosmos, Polkadot, Arbitrum, ZKSync, etc.
- Intuitive understanding of zero-knowledge proofs and their applications on blockchains.
Invested on some of projects mentioned above.
- Track record? Please don't ask.
Nerf Rust, Serde, Tokio, Tower, Hyper, HTTP
A HTTP request framework with
This is an experiment about cryptocurrency exchanges
API into composable
- Switchable HTTP client backends.
- Configurable rate limits (provided by
- Multiplexed HTTP clients (provided by
testing paired with
- Automatic throttling based on API responses with remaining rate limits (TODO).
WASynth(WIP) Rust, WebAssembly
Universal, language-agnostic instrumentation with WebAssembly.
- A standalone WebAssembly parser/generator implementation.
- Enter/leave hook injection into WASM binary functions.
- cargo fuzz-powered fuzzing.
Introduction to NIZK Proofs
Rust 비동기 프로그래밍 (Asynchronous Rust Programming)
- Korean: Native
- English: Business level
The favicon of this page was generated using following graphics from Twitter Twemoji:
- Graphics Title: 1f4f0.svg
- Graphics Author: Copyright 2020 Twitter, Inc and other contributors (https://github.com/twitter/twemoji)
- Graphics Source: https://github.com/twitter/twemoji/blob/master/assets/svg/1f4f0.svg
- Graphics License: CC-BY 4.0 (https://creativecommons.org/licenses/by/4.0/)