M.S. Computer Science · AI Depth · UC San Diego
Isaac Ang
I work on
Machine learning, math, and biology — from ChemBERT embeddings over recipe space to HMMs over a terabyte of sequence. Rock climbing and piano when I'm not coding.
ChemBERT recipe search
ML Engineer, Boston Bioprocess
Autoencoder + RAG, SageMaker
Jan – Apr 2026
- Extract 95% of recipe data with API queries and annotations; serialize and tokenize formulas, properties, and metadata to generate embeddings with ChemBERT
- Fine-tune an autoencoder with EDA/PCA/K-Means, uncovering 10 clusters from thousands of recipes; integrate into a transformer-RL pipeline
- Deploy SageMaker endpoints; refactor the TypeScript–FastAPI–RDS connection; streamline CI/CD with Git
Mental health app
Project Manager
Flutter-Django, agentic chatbot
Aug 2024 – May 2026
- Spearhead a fully containerized Flutter–Django app with agentic (Graphiti), breathing, and journal features; guided Android emulation, persistent storage, and E2E encryption
- Curate stateful animations; deploy to Cloud Run / Compute Engine VPC; enable login; write Bash tooling
Genomic anomaly detection
Stoch. Alg. Dev., Mirarab Lab
JAX EM-HMM, 1TB sequences
Jun 2026 – Present
- Devised a novel framework for Phlag, replacing discrete gene trees with continuous site-pattern CASTER scores to cut noise and raise the statistical resolution of genomic anomalies
- Redesigned HMM emissions and Expectation-Maximization in JAX to support multivariate Gaussians
Adaptive scheduler
ACTION Lab, Prof. Jessie Chin
TS-Node-Firebase, ACT-R + LLM
Oct 2024 – Present
- Deployed a TypeScript–Node–Firebase app that schedules, chunks, and prioritizes student tasks using interleaving, importance, and performance strategies from the psychology literature
- Prompt engineered an LLM to generate Lisp chunks for the ACT-R cognitive architecture, showing promise on math problems
Credit card ETL
Data Analyst, Synchrony
MySQL pipeline, 10k apps mapped
May – Jul 2025
- Mapped 10k new credit card applications to unique merchant identifiers, ensuring 1-1 correspondence across data warehouses with distinct aggregations
- Implemented a secondary mapping at 96% accuracy, then reached 100% by amplifying an existing attribute for inference and sourcing new warehouse datasets
- Enabled downstream Tableau analytics and client storytelling
Glycosylation pathways
UL Intern, Gunawan Lab
scRNA-seq, UMAP, NIH grant
Jun 2022 – Aug 2023
- 3rd author, Cell and tissue-specific glycosylation pathways informed by single-cell transcriptomics; $0.5M NIH grant
- Visualized single-cell expression for 400 glycogenes across 15 tissues with UMAP, t-SNE, and PCA; applied denoising and normalization to improve data quality
Pharma database
Software Developer
React-Node, Cloud SQL
Jan – May 2025
- Implemented a React–Node app with CRUD, stored procedures, and transactions
- Pulled data from trial, pharmacy, and drug APIs; connected the app to Cloud SQL
Algal bloom dynamics
Dyn. Sys. Researcher, Cooney Lab
Novel diff-eq, first-author SIAM
Apr 2025 – Present
- Developed novel differential equations modeling algal bloom behavior under seasonal light, water depth, and specific growth; first-author paper to be submitted to SIAM
- Time-stepped with RK-4 in MATLAB; modularized the simulation to enable comparison with theory
EEG classifier
Algorithms Lead, Neurotech
PSD/STFT, K-Means, MNE
Jan – Dec 2025
- Implemented signal processing algorithms (PSD, STFT) and K-Means to classify EEG signals
- Correlated fast time-series data with various noise events using Python MNE; compressed video with OpenCV
Stock predictor
AIMI x-ray classifier
NLP intern
skills
PythonC/C++RustTypeScriptSQLPyTorchJAXDjangoNodeDockerGCP
courses
AlgorithmsTheory MLNum. lin. alg.Real analysisCombinatoricsDatabasesBioinformaticsNLP syntax