Hello, I'm

Sheng Dong

Computer Science at University of Oxford

Building performant systems and intelligent applications. Currently interning at Tencent WeChat, focused on frontend optimization and distributed systems.

Scroll to explore

About Me

I'm a final year MCompSci student at Somerville College, Oxford, passionate about building software that scales. My interests span systems programming, compilers, and machine learning — essentially anything that makes computers faster and smarter.

When I'm not optimizing spreadsheet algorithms at Tencent or building due diligence platforms in London, you'll find me on the running track (I was High School Run Club President, after all) or planning my next continental adventure.

Performance Speed-up
60→1ms
Filter Latency Reduction
Days→1hr
Due Diligence Workflow

Education

University of Oxford, Somerville College

MCompSci in Computer Science

Expected Jul 2027

Grade: First Class Honours (Expected)

A Levels: Computer Science (A*), Physics (A*), Mathematics (A*), Further Mathematics (A*), Chinese (A*)

Relevant Coursework

Data Structures & AlgorithmsParallel ComputingCompilersMachine Learning

Awards & Scholarships

  • • Distinction in Preliminary Examinations
  • • Somerville Scholarship and College Prize (awarded twice)

Work Experience

Tencent, WeChat Group

Software Engineering Intern

Guangzhou, CN

Jul 2025 – Sep 2025
  • Redesigned frontend filtering/sorting for WeChat's spreadsheet platform, reducing filter latency from 60ms to 1ms and sort from 150ms to 30ms
  • Developed patent-pending spreadsheet optimization algorithm with regression replay testing and comprehensive unit tests
  • Built TypeScript frontend using lazy computation to load only required cells, reducing memory accesses and cache misses
  • Optimized C++ backend by replacing shared pointers with raw pointers, doubling performance (~2× speed-up) in large-scale numerical operations
  • Created intelligent feedback analysis pipeline using K-Means clustering and LLM-based evaluator for adaptive hyperparameter tuning

TSL Projects Ltd

Software Engineering Intern

London, UK

Jun 2024 – Aug 2024
  • Built end-to-end due diligence platform using React, Django and Python, reducing workflows from days to 1–2 hours
  • Migrated legacy frontend from Vue to React using Hooks and Context, improving modularity and reducing code overhead
  • Implemented Redis-backed caching to optimize backend request handling and permission checks
  • Set up automated CI/CD pipelines on Azure, cutting deployment effort to minutes

Technical Projects

Type-Safe Communication Modelling for Distributed Microservice Engine

CurrentApr 2025 – Present

Designed communication protocols for distributed microservices using Multi-Party Session Types, defining global message flows that guarantee deadlock-freedom and preserve liveness across coordinated computations.

Multi-Party Session TypesGuarded Attribute GrammarsType Theory

ILookGood App

Apr 2025

Led a five-person team through requirements analysis, architectural planning and sprint management, delivering a multi-modal AI outfit-analysis app from concept to MVP within eight weeks.

  • Integrated multi-modal LLM pipeline for outfit parsing, colour analysis and style scoring, reaching over 90% prediction consistency
  • Enabled smooth image upload, efficient model invocation and consistent interface rendering under 300ms
FlutterMulti-modal LLMComputer VisionPython

Skills & Expertise

Languages

CPythonJavaScriptTypeScriptSQLSCSS

Frameworks & Libraries

Next.jsReactFlutterDjangoTailwind CSS

Tools & Infrastructure

GitJestRedisWebpackGDBPostgreSQLAzure

Spoken Languages

English (Fluent)Mandarin (Fluent)

Beyond Code

🏃

Running

Former High School Run Club President. Believe me, the morning training sessions were worth it for the medals.

✈️

Travelling

Completed a Continental Asia Tour. There is something about experiencing different cultures that reshapes how you think about software and design.

Let's Connect

Currently open to new opportunities. Feel free to reach out if you'd like to discuss technology, running routes, or travel recommendations.

LinkedInGitHub
Oxford, UK