Here's a collection of projects I've worked on, from systems and web development to deep learning and compilers.
MLflow PyTorch CI/CD GitHub Actions Docker Machine Learning May 2026 - June 2026
I integrated a PyTorch training workflow with full MLflow lifecycle management, covering experiment tracking, the model registry, and automated promotion. The goal was to eliminate the manual overhead of comparing model iterations and tracking which version was deployed, replacing it with a system that handles versioning and promotion automatically based on measured performance.
Machine Learning Bioinformatics Python scikit-learn Data Pipelines April 2026 - June 2026
This project classified COVID-19 severity (Severe vs. Non-severe) from mass spectrometry proteomic data using a leakage-free machine learning pipeline. The dataset started at 101,461 peptide variants across 268 columns and was reduced to a 43 patient by 200 peptide matrix through a structured preprocessing pipeline before any model saw the data.
AI Agents LLM Systems GitHub Actions Python Claude SDK CI/CD May 2026 - June 2026
I built a multi-agent code review system on the Anthropic Claude SDK that integrates directly with GitHub Actions to autonomously analyze every pull request for security vulnerabilities and code quality issues, then posts structured findings as PR comments with zero manual intervention.
AI Agents LLM Systems Benchmarking Python OpenAI Debugging January 2026 - March 2026
I built a custom LLM debugging agent framework from scratch, designed to systematically benchmark how different agentic workflow protocols, models, and fix strategies perform across a structured set of real debugging tasks. The goal was to move beyond anecdotal comparisons and produce reproducible, structured evidence for which configurations actually work and at what cost.
Web Development Node.js PostgreSQL Docker GitHub Actions Leadership September 2025 - December 2025
Conductor is a lightweight platform helping instructors, TAs, and students manage course activities including attendance, scheduling, and group tracking. I was team lead for this project, managing a team of 12 split across frontend, backend, database, and testing teams. The application is deployed live on AWS EC2 and built with Node.js, Express, and PostgreSQL on the backend, with a vanilla HTML, CSS, and JavaScript frontend.
AI Agents LLM Systems Memory Systems Redis Vector Databases Docker AWS September 2025 - December 2025
I designed and implemented a context-aware AI agent system that supports both short-term and long-term memory across multiple users and sessions. The goal of this project was to move beyond stateless chatbots by building an agent that can remember prior conversations, retain user preferences, and reason coherently across sessions, while remaining scalable and isolated at the systems level.
Python OpenAI API ServiceNow Automation September 2025 - October 2025
In this project, I created an AI-driven incident classification agent that integrates ServiceNow with the OpenAI API. The tool automatically categorizes incidents, helping IT teams reduce repetitive work and maintain consistent labeling. Below is the full workflow with examples.
AWS DevOps CI/CD Security GitHub Actions Jekyll August 2025 - September 2025
In this project, I deployed melvyn-tan.com as a secure static website hosted on AWS. The site uses a private S3 bucket with CloudFront as the CDN, Route 53 for DNS, and ACM certificates for HTTPS. I set up logging, CloudTrail, and budgets for monitoring, and automated deployments first with AWS CodePipeline/CodeBuild before migrating to GitHub Actions with OIDC for long-term use.
Web Development React TypeScript CI/CD GitHub Actions Leadership Educational Technology April 2025 - Current
OCCTIVE (Online Computing-Concepts Toolkit of Interdisciplinary Videos for Education) is an NSF-funded platform serving 1,000+ students across 15+ universities in the U.S., aimed at helping non-CS faculty introduce foundational computing concepts to students in non-computing courses. I am the lead developer for this React and TypeScript web application, managing both the front-end and back-end teams while presenting progress and pitching designs to professors across the country in bi-weekly meetings.
AWS Lambda Amazon ECR Docker Serverless July 2025 - August 2025
I worked on a containerized application to get familiar with running containers in AWS Lambda. The project involved building a Docker image, pushing it to Amazon ECR, and deploying it as a Lambda function using Lambda’s container image support.
AWS Cloud Computing Data Engineering Serverless Python June 2025 - August 2025
I built a clickstream ingestion pipeline on AWS to familiarize myself with real-time data processing and analytics. The project connected API Gateway, Kinesis Data Firehose, Lambda, S3, Athena, and QuickSight into a working architecture for transforming, storing, and visualizing clickstream data.
AWS Cloud Computing Event-Driven Architecture Serverless Python June 2025 - August 2025
I built an event-driven architecture on AWS to familiarize myself with core services and how they integrate to form scalable, serverless applications. The system simulated an order processing pipeline with notifications.
AWS Kubernetes Amazon EKS Docker DevOps July 2025 - August 2025
I worked on a containerized directory application to get familiar with Amazon EKS and Kubernetes on AWS. The project involved building Docker images, pushing them to Amazon ECR, configuring IAM policies, and applying Kubernetes manifests to deploy backend and frontend services on EKS.
Parallel Computing Automatic Differentiation OpenMP MPI Compiler Design March 2025 - June 2025
I extended a research compiler framework for automatic differentiation (AD) to support distributed execution across multiple cores using OpenMP and MPI. The goal was to allow users to define any numerical function, regardless of input dimensionality, and automatically generate C code that computes derivatives in parallel.
AWS Distributed Systems etcd Consensus DevOps June 2025 - July 2025
I set up a three-node etcd cluster on AWS EC2 to gain first-hand experience with distributed consensus, leader election, and fault tolerance. The project involved launching virtual machines, configuring firewall rules, installing etcd, and experimenting with cluster behavior under simulated failure conditions.
Deep Learning Music Generation Music Transformer Sequence Modeling PyTorch March 2025 - June 2025
I trained a Music Transformer from scratch to generate symbolic music using the POP909 dataset. The goal was to build a model capable of composing music by predicting the next token in a sequence of encoded musical events.
GPU Programming Triton Matrix Multiplication Deep Learning Optimization CUDA February 2025 - March 2025
In this project, I wrote a custom matrix multiplication kernel using Triton that fuses three operations—A @ B + C, followed by a ReLU activation—into a single GPU kernel. The goal was to outperform PyTorch’s high-performance cuBLAS backend by taking advantage of kernel fusion and manual tiling. My Triton implementation achieved a 1.4× speedup over the equivalent PyTorch function on fp16 tensors.
Distributed Systems Consistent Hashing gRPC Go SQLite Video Streaming April 2025 - June 2025
Distributed content delivery is critical to scalable systems like YouTube, especially when reliability and fault tolerance are required. TritonTube simulates this at a systems level, allowing videos to be uploaded, segmented using MPEG-DASH, and distributed across a network of storage nodes using consistent hashing. As part of a networking course, I built this application from the ground up in Go and achieving full functionality.
Wireless Networks Internet of Things LoRa Energy Modeling Systems Design Cost Analysis January 2025 - March 2025
As part of my Wireless Network Systems class, I designed a smart farming solution for a 40-acre plot of land in Malaysia. The goal was to modernize traditional farming methods using wireless sensors to monitor soil moisture and nutrient levels (nitrogen, phosphorus, potassium). Each sensor transmits 11 bytes of data twice a day—at 6am and 6pm—and undergoes firmware updates three times a year.
Recommender Systems Collaborative Filtering Matrix Factorization Factorization Machines Data Processing October 2024 - January 2025
Recommender systems are essential for platforms like Yelp, where users rely on personalized suggestions. In this project, I built a system to predict how a user would rate a business they haven’t visited, using real-world Yelp data.
GPT-2 NLP Summarization Token Optimization Prompt Engineering Deep Learning October 2024 - January 2025
Large language models like GPT-2 are powerful for summarization but can be expensive to use, especially when deployed via APIs where cost is based on token count. This project focused on reducing input token usage while maintaining—or even improving—summarization quality.
Parallel Computing Matrix Multiplication CUDA GPU Computing Shared Memory Tiling Warp Utilization October 2024 - November 2024
Matrix multiplication is a fundamental operation in high-performance computing, and optimizing it for GPUs requires an in-depth understanding of memory hierarchy, parallel execution, and warp scheduling. In this project, I implemented a highly optimized CUDA-based matrix multiplication kernel, leveraging 2D tiling, shared memory, and interleaved computation to maximize efficiency.
GPT NLP Transformers Attention Mechanism Language Modeling Deep Learning November 2024 - January 2025
Transformers are the foundation of modern NLP models, and for this project, I developed a GPT-style model from the ground up. My implementation included a Transformer Encoder for classifying political speeches and a Transformer Decoder for generating text in an autoregressive manner, following the same structure used in GPT models.
Deep Learning Machine Learning PyTorch Research Image Processing March 2024 - September 2024
As a research assistant in the Machine Learning, Perception, and Cognition Lab at UCSD, I worked under the supervision of PhD students and Professor Zhouwen Tu. My role was that of a research engineer, focused on making experiments run reliably rather than driving research direction. I implemented and adapted PyTorch-based training pipelines from published papers, validated results within ±3% of reported benchmarks, and built experiment tracking and pipeline automation to reduce manual overhead across training, evaluation, and validation.
Machine Learning Unsupervised Machine Learning Python Scikit-Learn March 2024 - June 2024
In this project, I classified songs into distinct genres using the Spotify 1.2M+ Songs dataset. The goal was to group songs based on shared musical features such as danceability, tempo, acousticness, and valence. I hypothesized that songs within the same genre would exhibit similar patterns in these variables.
Backend Development MongoDB Mongoose RESTful APIs Express.js Authentication Web Development Jira May 2023 - July 2023
I worked on RoomEase, a platform that simplifies room rentals, where I focused on backend development. I designed and implemented data schemas using Mongoose and managed the database with MongoDB to store and retrieve user and room information. I also built RESTful APIs with Express.js to allow for efficient interaction between the frontend and backend.
Web Development React TypeScript Accessibility WCAG SEO GitHub Node.js Yarn Netlify Agile April 2022 - Current
As a web developer for the ComputingPaths project, I worked to enhance both the functionality and accessibility of a React-based web application. This platform is designed to inform prospective UCSD students about the various engineering majors offered at the university. My role involved collaborating with other developers, improving the user experience, and creating new content.
