Dennis Liang profile picture

Hello, I'm

Dennis Liang

Software Engineer & MEng Student

My LinkedIn profile My GitHub profile
Scroll to About

Get To Know More

About Me

Profile picture
Experience icon

Experience

7+ years
Software Development
Education icon

Education

MEng EECS at UC Berkeley
BS Computer Science at UC San Diego

Engineer with a strong interest in visual computing and computer graphics, combining technical depth with creative problem-solving. Experienced in developing interactive AR/VR applications in Unity, implementing computer vision and rendering techniques, and exploring how visual systems can enhance human understanding. Passionate about building efficient, visually compelling software that connects perception, visualization, and technology.

My

Experience

Software Development Engineer Intern

Amazon • Seattle, WA

  • Designed and implemented a targeting service to segment sellers, aiming to reduce 45,000+ annual hours spent on manual segment list creation
  • Built and deployed scalable backend systems using AWS services including OpenSearch Service, DynamoDB, S3, ECS, and EC2.
  • Collaborated cross-functionally to integrate the service into internal workflows while maintaining security and operational best practices.

Software Engineering Intern

Robolink • San Diego, CA

  • Launched and enhanced Robolink’s Web Python editor using React Pyodide: Identifying and resolving critical issues, improving functionality and user experience, contributing to a successful launch.
  • Tested and implemented improvements for drone systems: Conducting comprehensive testing on drones, fixing software bugs, and developing a new dashboard system to streamline classroom learning management and enhance educational experiences.
  • Streamlined the development pipeline from coding and testing to deploying new features, ensuring consistent performance improvements and maintaining high-quality standards across the software.

Undergraduate Research Assistant

UC San Diego Center for Visual Computing • San Diego, CA

  • Assisted in visual reconstruction research by developing/optimizing imaging algorithms and testing FLIR Blackfly S camera software, significantly improving image quality.
  • Calibrated advanced imaging equipment and managed complex setups to ensure experimental accuracy and efficient data collection.

Browse My Recent

Projects

UCSD Cycling Team Website logo

SpatialFind

VR Image Retrieval System

  • Engineered an end-to-end AI-powered image retrieval system using Python (FastAPI), CLIP ViT-B/32, and FAISS vector search, implementing a 4-stage progressive narrowing pipeline (90 → 30 → 10 → 1) with softmax temperature-scaled re-ranking over a 25,000-image corpus
  • Built a cross-platform XR client in Unity (C#) with asynchronous HTTP communication via UnityWebRequest coroutines, supporting seamless deployment across PC simulation and HTC Vive Focus Vision (Android/ADB) by swapping a single server endpoint
  • Optimized GPU inference pipeline on a 4GB VRAM-constrained RTX 3050 Ti using batched CUDA processing and memory-aware batch sizing, reducing embedding generation time across the full dataset corpus
GitHub
UCSD Cycling Team Website logo

Virtual Reality Based Learning Environment

A VR STEM learning platform built for Meta Quest 3S devices

  • Contributed to the development of immersive, interactive VR modules for STEM education.
  • Improved scene hierarchy and code modularity to support multi-topic lesson extensions.
  • Collaborated with designers to refine UI/UX for classroom deployment readiness.
GitHub
UCSD Cycling Team Website logo

AWS CDK Linter

Static analysis tool for detecting AWS CDK misconfigurations before deployment

  • Contributed to a team-built Python static analysis tool for AWS CDK applications, conducting contextual inquiry and need-finding to identify three key misconfiguration categories: single-resource errors, cross-resource connection issues, and application-level dependency problems
  • Designed a modular rule engine where each check operates on structured data models derived from TypeScript ASTs or compiled CloudFormation YAML templates, treating code as structured data rather than raw text
  • Worked across parallel parsing pipelines (TypeScript and CloudFormation) with a unified interface enabling independent rule authorship across team workstreams
GitHub
VibeRight Project Image

VibeRight

An AI-powered code understanding tool that tests and strengthens a developer’s grasp of their own system before they ship

  • AI-Generated System Questions – Analyzes your codebase and generates targeted questions to test real understanding
  • Adaptive Feedback Loop – Evaluates your answers and drills deeper where gaps exist.
  • Ship with Confidence – Ensures you can explain your architecture, logic, and edge cases before deployment.
GitHub
Ferris Project Image

Ferris - AI Travel Agent

A voice-first AI travel planner with natural language itinerary generation

  • Developed core UI components for trip planning and natural-language interaction.
  • Integrated Gemini and Deepgram API for personalized destination suggestions and real-time itinerary updates.
  • Optimized FastAPI backend for faster query response times.
GitHub
Volumetric Path Tracer Image

Volumetric Path Tracer

A physically based renderer simulating participating media and volumetric effects

  • Implemented Monte Carlo–based volumetric light transport for realistic fog and scattering effects.
  • Designed acceleration structures to improve rendering performance and reduce noise.
  • Validated renderer results using Cornell box and custom participating media scenes.
Project Page
Ray Tracer Image

Ray Tracer

A CPU-based ray tracer supporting global illumination features

  • Built a physically-based ray tracing engine supporting reflections, refractions, and shadows.
  • Implemented BVH acceleration to achieve real-time performance on test scenes.
  • Added tone mapping and anti-aliasing to improve visual fidelity and rendering quality.
Zooseeker logo

Zoo Seeker

An Android navigation app for optimized zoo route planning

  • Led Android app development to plan optimized zoo routes using Dijkstra’s algorithm.
  • Designed intuitive UI for animal selection, route visualization, and real-time navigation.
  • Managed data persistence and location updates using MVVM and Android Jetpack components.
GitHub
Say-It Chat Assistant preview

Say‑It Chat Assistant

Voice‑enabled chat assistant with a clean, accessible interface for quick interactions

  • Managed conversation state and accessibility features for seamless user experience.
  • Owned core UI flows and state management.
  • Focused on usability and responsive behavior.
GitHub
Cycling Points Calculator logo

Cycling Points Calculator

Tool to compute and visualize USA Cycling points for fast, accurate lookups

  • Developed a Python-based web app to calculate USA Cycling upgrade points instantly.
  • Visualized race results and point thresholds with interactive charting libraries.
  • Automated data input and verification to minimize user error and manual lookup time.
UCSD Cycling Team Website logo

UCSD Cycling Team Website

Improved the team site structure and content for better readability and updates

  • Modernized the team website with updated race reports, officer profiles, and sponsor pages.
  • Improved site performance and responsiveness with streamlined CSS and layout refactor.
  • Collaborated via GitHub to coordinate open-source contributions and content updates.
GitHub

Get in Touch

Contact Me

Email icon

dennisliang01@gmail.com

LinkedIn icon

LinkedIn