Skip to content

Latest commit

 

History

History
182 lines (127 loc) · 7.08 KB

resume-all-in-one.md

File metadata and controls

182 lines (127 loc) · 7.08 KB
Raw

YU WENYA - Resume -ALL In One

::: left

icon:blog Singapore

icon:github https://github.com/yushen611

:::

::: right

icon:email YU0001YA@e.ntu.edu.sg

icon:phone +65 80674181

:::

Education Background

:::left Nanyang Technological University - Computer Science (MSC-blockchain) ::: :::right 2023.08-2024.07 :::

  • Research in NTU CSLab Monitoring and Alerting of Underlying Trading Risks by Large Language Models(LLMS)

:::left Hefei University of Technology - Software Engineering (B.E.) ::: :::right 2019.09-2023.06 :::

  • Grades and Ranking: 91.02/100 (Ranking 5/169)
  • Awards and Honors: National Scholarship, First-Class University Scholarship, Outstanding Student Award, Third Prize in National Computer Design Competition

Internship Experience

:::left Tiktok(Singapore) - Golang Backend ::: :::right 2023.12-2024.06 :::

FaaS AWS Lambda workflow

  • Implemented over 10 atomic services utilizing Function as a Service (FaaS) tools similar to AWS Lambda.
  • Developed self-recover workflows for ingress and node resource recovery, enhancing the uptime of push and pull stream nodes to over 90%.

:::left USTC GUOCHUANG CLOUD Technology Co.,Ltd - Java Developer ::: :::right 2021.12-2022.02 :::

Spring Boot OCR Redis

  • Achievement: Obtained the software copyright for "Intelligent Enterprise Financial Reimbursement System" (First Author).
  • VAT Invoice OCR Recognition: Implemented invoice OCR recognition and automatic entry using YOLOv3 + CRNN + CTC.
  • Authentication and Authorization: Utilized JWT for authentication and authorization, achieving stateless authentication on the server side.
  • Preventing Cache Avalanche: Implemented distributed mutual exclusion locks using Redis to prevent cache avalanches.

Project Experience-Backend

:::left Online Video Playback System - Golang Backend Developer ::: :::right 2023.03-2023.06 :::

Gin GPRC Consul Redis dcoker CI/CD Kubernetes Prometheus

  • Overview: Developed an online video streaming platform with gin and grpc, utilizing microservices and distributed architecture for high QPS and stability. Implemented logging, monitoring, and CI/CD to significantly improve performance and efficiency.
  • System Metrics Monitoring and Alerting: Monitored server status using Prometheus, Grafana, and Altermanager. After Post-Load Testing through Jmeter, performance bottleneck analysis is employed with ELK log recording and PPROF flame graphs for performance bottleneck analysis.
  • High Concurrency and High Performance: Achieved an interface QPS of over 1500 by using Redis, establishing DB Index and Read-write separation, Server Scale Out with Kubernetes, and a pagination query to optimize the slow-query video list interface.
  • CI/CD: Compressing the entire process from Code Changes to Deployment into a 2-minute timeline by implementing CI/CD using Docker and Github Action.

Project Experience-AI

:::left Blockchain Transaction Risk Monitoring Based on LLM - NTU CSLAB ::: :::right 2023.08-2023.12 ::: LLM LLaMA2 Fine-tuning

  • Summary: Fine-tuned LLaMA2 to create a large model for EVM transaction risk detection.
  • Large Language Model Fine-tuning: By fine-tuning a large language model, we enhanced its semantic mapping capabilities for transaction texts, enabling accurate transaction risk classification with an accuracy of 80%.
  • Prompt Optimization: Improved the model's risk identification accuracy further through techniques like Gradient-based Optimization of Text (GOT) and other prompt optimization methods.

:::left AI Text-to-Video Generator - Virtual Reality AI-Lab In HFUT ::: :::right 2022.03-2022.12 ::: TransCGAN Text-to-Video

  • Summary: Enhanced extraction of spatiotemporal sequence information: The published TransCGAN model successfully merges CGAN with Transformer, capable of instantly generating customized human action videos based on textual commands.
  • Paper： "TransCGan-based human motion generator" (First Author)

:::left Web SQL Injection ML-Detector - AI-Lab In HFUT ::: :::right 2023.03-2023.06 :::

SQL Injection machine learing nlp

  • Summary: Developed a machine learning-based SQL injection intelligent detector with a detection accuracy of 97%.
  • Word Embedding Optimization: Utilized the word2vec algorithm integrated with positional encoding and combined with a one-dimensional convolutional pooling layer to efficiently vectorize and denoise SQL query strings.
  • Model Selection: After training various machine learning models and comparing different metrics, the Random Forest model, which showed the highest accuracy on our dataset, was selected.
  • Dataset: Independently collected by us, it is currently the largest SQL Injection (SQLI) dataset available on the internet.

Project Experience-BlockChain

:::left Open-Indexer For EVM Inscription - NTU ::: :::right 2024.01-2024.02 :::

EVM ERC-7583 inscription Indexer

  • Summary: Implemented an EVM inscription system and an open-source indexer by implementing the ERC-7583 protocol.
  • lowest gas fee: By storing NFT / FT in transaction events, the cost for each minting is brought close to the minimum transaction fees
  • fast performance: Improved the indexer's performance with caching and incremental reading optimizations, boosting the response time from 13 seconds to just 1 second

Key Skills-SOFTWARE

  • Summary: Golang Backend, Distributed System, DevOps, High-performance Software
  • Language skills: IELTS score of 6.5
  • Development Language: Go, Python, Java, C++
  • Development Framework: GIN, GORM, Spring Cloud, Spring Boot
  • Database: MySQL, ClickHouse
  • Middleware: Consul, GRPC, Redis, RocketMQ, Prometheus, Grafana, Altermanager, EKL
  • Operations: Linux, CI/CD, Docker, Kubernetes

Key Skills-AI

  • Summary: DL & ML, NLP, Fine-Tuning LLM
  • Language skills: IELTS score of 6.5
  • Development Language: Go, Python, Java, C++
  • AI: Pytorch, Opencv, LLAMA2, Transformer, GAN

Key Skills-BC

  • Summary: BlockChain, Golang Backend, Distributed System, High-performance Software
  • Language skills: IELTS score of 6.5
  • Development Language: Solidity, Go, Python, Java, C++, JavaScript
  • Development Framework: Web3.js, HardHat, GIN, GORM, Spring Cloud, Spring Boot
  • Database: MySQL, ClickHouse
  • Middleware: IPFS, Consul, GRPC, Redis, RocketMQ, Prometheus, Grafana, Altermanager, EKL
  • Operations: Linux, CI/CD, Docker, Kubernetes

Thesis, Patent, and Software Copyright

  • Thesis: "TransCGAN-based Human Motion Generator," International Conference on Computer Information Science and Artificial Intelligence. First Author, 08.2022
  • Software Copyright (SC): "Intelligent Financial Reimbursement System," No. 2022SRO576160. First Author, 03.2022
  • SC: "Beichite Cloud Customized Printing System," No. 2022SRO758646. First Author, 03.2022
  • Patent: "A Wireless Torque Sensor," No.202210304204.9. Second Inventor, 06.2022
  • Patent: "A Walking Aid Stick," No. 202010166261.6. Second Inventor, 05.2020