B.S. Computer Science, UCLA · Expected 2029

Andrew Woods

I build ML systems and then measure whether they actually work — agentic pipelines, retrieval, evaluation. Usually the measurement is the interesting part: the last one found that the reranking step almost everyone adds by default lowered retrieval quality on two of three datasets.

Most recently: Software Engineer Intern at OmicsAI Labs, where I built agentic multi-agent pipelines for causal inference over multi-omics data.

Projects → Publications →


Experience

Software Engineer Intern

OmicsAI Labs June 2026 – Aug. 2026

Remote

  • Architected a React frontend backed by FastAPI and a Gradio AI interface over an agentic multi-omics system.
  • Deployed containerized services to AWS EC2 (Docker Compose, Caddy, HTTPS, subpath routing), shipping omicsai.org/agentic-causal to the lab's demo platform serving papers under review at Oxford Bioinformatics.
  • Developed LangGraph_Orchestrator, a LangGraph and Streamlit service coordinating multi-agent workflows across the platform.

AI Research Intern

Georgetown University Feb. 2024 – Nov. 2025

Dept. of Biostatistics, Bioinformatics & Biomathematics · Washington, DC

  • Designed a plan–act–verify architecture (LangGraph, GPT-4o) to evaluate whether agentic orchestration can replace hand-scripted RNA-seq pipelines, adapting preprocessing and embedding choices to the biological question.
  • Formalized quantitative acceptance criteria — variance explained, resampling stability, cluster coherence, pathway concordance — as automated verification gates governing agent decisions.
  • Encoded curated pathway structure (KEGG, Reactome, GO:BP, WikiPathways) as priors via gene–pathway incidence and Jaccard co-membership adjacency over DESeq2 variance-stabilized expression.
  • Designed the ablation protocol benchmarking pathway-weighted embeddings against plain PCA and direct differential-expression baselines, with per-run artifact logging for reproducibility.

Education

B.S. Computer Science

University of California, Los Angeles Expected 2029

Los Angeles, CA

Relevant coursework: Multivariable Calculus, Differential Equations, Linear Algebra, Discrete Mathematics, Computer Programming 3: Advanced Topics in Computer Science.


Technical skills

Languages

  • Python
  • Java
  • JavaScript
  • SQL
  • HTML/CSS

AI & ML

  • LLM applications
  • RAG
  • Agentic AI
  • Multi-agent systems
  • LangGraph
  • OpenAI/GPT APIs
  • Embeddings
  • Weighted PCA
  • Information retrieval
  • BM25
  • FAISS
  • Cross-encoder reranking
  • nDCG/Recall evaluation

Frameworks & infrastructure

  • React
  • FastAPI
  • Streamlit
  • Gradio
  • REST APIs
  • AWS EC2
  • Docker Compose
  • Caddy
  • Linux deployment

Bioinformatics

  • DESeq2
  • g:Profiler
  • KEGG
  • Reactome
  • Gene Ontology (GO:BP)
  • WikiPathways
  • Multi-omics analysis

Tools

  • Git
  • GitHub
  • GitHub Actions
  • pytest
  • Vector search
  • VS Code
  • Conda
  • Linux/Terminal

Contact

Fastest way to reach me is email.

andrew.woods.jason@gmail.com

If your browser has no mail app set up, that address is selectable — just copy it.

GitHub LinkedIn