Resume

Andrew Hua

Junior at Carnegie Mellon University studying AI and Applied Math

Hi, I'm Andrew! I'm from Toronto and now live in Pittsburgh, where I study AI and applied math at Carnegie Mellon. I like building full-stack applications and designing data pipelines for AI systems.

Outside of class and work, I enjoy bouldering, watching soccer, and trying new restaurants. I'm always open to trade food recs!

Technical skills

Languages

  • Python
  • SQL
  • TypeScript
  • Java
  • C++

ML and AI

  • PyTorch
  • NumPy
  • Pandas
  • RAG
  • MCP
  • LLMs

Data and infra

  • Databricks
  • MongoDB
  • Postgres
  • Docker
  • Kubernetes
  • Git
  • Linux

Web

  • Next.js
  • React
  • Angular

Experience

S&P Global

Software Engineer Intern

Jun 2026 – Aug 2026

New York, NY

  • Built a multi-agent workflow (15+ agents) that produces business insights and data visualizations on the reliability of High Value Ratings applications, used by 2M+ external users and 40k+ employees. Cut the division-wide application error rate by 6%.
  • Built MCP tools, servers, and conversational agents that let the CTO and senior leadership search 3M+ error logs, latency events, and problem tickets with natural-language queries, replacing manual pulls from Databricks and Splunk Observability Cloud.
  • Built a repository scanner using an internal LLM and a RAG pipeline over 600+ pages of Oracle, Postgres, and MongoDB best-practice documentation, flagging dangerous, inefficient, and unreliable database queries before deployment.

AnthrAI

Software Engineer Intern

Dec 2025 – Jun 2026

San Francisco, CA

  • Built AI web-navigation agents that simulate 6 configurable user personas with over 90% behavioral similarity to real users, enabling automated hypothesis testing on third-party web interfaces.
  • Shipped an end-to-end UI/UX testing platform (Next.js, Vercel) that orchestrated 12k+ automated agents across 400+ A/B and hypothesis-test batches for 1,000+ users.
  • Developed KPI and performance metrics in Python and SQL across 1,200+ simulation batches, cutting simulation failures by 25% and agent navigation steps and API costs by 40%.

Education

Carnegie Mellon University

B.S. in Computational and Applied Mathematics

B.S. in Artificial Intelligence

Aug 2024 – May 2028

Pittsburgh, PA

GPA
3.8 / 4.0
Honors
Dean’s List, High Honors: Fall 2024, Spring 2025, Fall 2025, Spring 2026
Coursework
Foundations of Software Engineering, Artificial Intelligence, Machine Learning, Data Structures and Algorithms, Autonomous Agents, Functional Programming, Theoretical Computation

Projects

Dynamic memory allocator

C

  • Implemented malloc, free, realloc, and calloc, evolving the free-block structure from an implicit list to a segregated free list indexed by size class. Throughput rose 65× to 6,500+ KOPS and memory utilization reached 74% through block coalescing and footer elimination.
  • Wrote a heap consistency checker that validates block alignment, header integrity, and free-list correctness to catch memory corruption systematically.

RNN language model with self-attention

Python

  • Built an RNN language model with self-attention, a custom tokenizer, and greedy and stochastic decoding, trained on 100K short stories (10K held out) across 20 configurations on cloud GPUs.
  • Reached a validation loss of 2.30 (about 10 perplexity), tracking and plotting evaluation metrics with NumPy and Matplotlib.

Awards

  • Champion

    Canadian Intermediate Mathematics Competition

    Top individual score among 9,000 contestants worldwide

  • 2nd place

    Harvard-MIT Mathematics Tournament

  • 4× qualifier

    American Invitational Mathematics Examination (AIME)

  • Global Top Scorer

    PhysicsBowl Contest

  • Qualifier

    Canadian Mathematical Olympiad Repechage

    Top 100 score on the Canadian Open Mathematics Challenge

mainPittsburgh, PAMarkdownUTF-8