Software Engineer
Based in Seoul š°š·
updated: @August 13, 2022
Stack
Kubernetes, Argo, Faiss, Milvus, Airflow, ELK, Redis, Kafka, Hadoop, HBase, TiDB @For the details
š» Selected Projects
GPU Cloud for Deep-Learning Inference
- Apache Mesos - Docker, Ansible, NVML - Python, Bash
GPU Cloud is a platform for deploying Deep-Learning inference. Based on information of GPU utilization with NVML(NVIDIA Management Library), developer and researcher could deploy multiple applications on a single GPU with scheduling algorithm that has made improve GPU server utilization.
Machine Learning Platform
- Apache Mesos - Docker, Ansible, Vue - Python, Go, Bash
Machine Learning Platform helps researchers train Tensorflow or PyTorch model experiments easily with command line interface and Web UI. The platform also has made improve GPU server utilization based on GPU Cloud.
Feature Vector Similarity based Search Engine
- Kubernetes, Docker - Hadoop, HBase, ELK, Promethues, Grafana - Go, Scala, C/C++
Image Search Engine uses image feature vectors, HNSW(Hierarchical Navigable Small World) and PQ + OPQ(Optimized Product Quantization) for Approximate KNN Modeling. It splits data into each shards for indexing through many nodes. If the feature vector has a large dimension, quantizing through OPQ, the nodes are able to store the larger data on memory.
Face Recognition SDK for mobile
- CoreML - OpenCV - Python, Swift, Objective-C, C/C++
Face Recognition SDK for mobile environments starts with face image stream input then works through on pipeline including deep-learning models of face detection, alignment for facial landmarks, feature extraction and recognition. With the pipeline based on C/CPP, the SDK supports Android, iOS, Python interface with CoreML and NCNN frameworks both.
E-Commerce Recommendation System
- Kubernetes, Argo Workflow - Hadoop, HBase, Redis, Spark, ELK - Python, C/C++
E-Commerce Recommendation System is the first project I've taken on Kakao Corp. Therefore I got through a lot of moving the applications and batch jobs on bare-metal servers to Apache Mesos framework and then Mesos to Kubernetes. This project also gave me a lot of valuable lessons about huge data handling systems like batch jobs, Hadoop file system, HBase, Redis, Spark and real time API deployments.
šØš»āš» Work Experience
2022
Machine Learning Engineer @Daangn , Seoul, S.Korea
2021-2022
2019-2021
Software Engineer @Kakao Enterprise , Bundang, S.Korea
Kakao Enterprise is a company that provides AI services to companies spun off from Kakao's AI Lab department. David division, Vision Team, AI Lab at Kakao Enterprise is for massive image embedding vector search and facial recognition.
- Image Feature Vector Similarity Search Engine (Build own system)
- Cross mobile platform Face Recognition SDK (supporting iOS, Android)
- ML Pipeline for Celebrity Face Recognition & Identifier
2017-2019
Software Engineer @Kakao , Jeju, S.Korea
Multimedia Processing division, AI Lab at Kakao is for processing multimedia like video, image, music and voice with deep-learning.
- GPU Cloud for Deep-Learning Inference
- Machine Learning Platform for Traning Neural Networks model
- E-Commerse CF & CB Recommendation System
2015-2016
Software Engineer & Co Founder @RISE (Startup Company), Pohang, S.Korea
- DIY Parts Recommendation System for proto-typing device.
š Honors & Awards
2020
Excellence Award, @OSS , Seoul, S.Korea
Open Source Software Contribution for @RustPython
2019
3rd Place, @Kakao, Pangyo, S.Korea
Kakao In-house 24K Hackerton
- Image Search with Text on Kakaotalk Messenger
- Chrome extension, automatically generating image descriptions for a blind to harness from VoiceOver
2015
Bronze Prize, @ICISTS-KAIST, Daejeon, S.Korea
ICISTS-KAIST Team Project Award, Bronze Prize in All, Silver Prize in Robotics
šļø Certificates
2021
2019
2018
Iām available for interesting projects. See me in action.
Thanks for reading! š