Hi, I'm Henry Hong, Software Engineer and ML Researcher at UC Berkeley

Software Development Engineer Intern

Amazon

• Designed and deployed a multi-agent pipeline leveraging AWS Bedrock to detect anomalous tax rates in 10M+ invoices/year
• Built a RAG-based database for tax rate recovery and a tax-line extraction agent in SageMaker to parse invoices into structured schemas
• Created retrieval and web-scraping agents to cross-verify tax data against state repositories, cutting manual review 90% and preventing $2.27M in overcharges
• Benchmarked AWS Strands SDK vs. the Bedrock API, finding Bedrock 2.4x faster per invoice with 3.2x fewer tokens

• Tech: AWS Bedrock • SageMaker • RAG • Multi-Agent Systems • Python
• Link: Amazon
• Date: Summer 2025

Data 100 Course Staff

UC Berkeley CDSS

• Taught and supported 1,000+ students through weekly sections, exam prep, and daily course website updates across semesters
• Led discussion sections and authored lab assignments and exam prep materials used course-wide
• Held office hours guiding students through pandas, EDA, data visualization, and ML techniques like linear & logistic regression

• Tech: Python • Pandas • NumPy • Jupyter • Data Science • Machine Learning
• Link: Data 100
• Date: Spring 2025 – Spring 2026

Computational Neuroscience Researcher

UCSF

• Conducted research at UCSF to optimize seizure detection models for personalized clinical care
• Analyzed MEG/EEG data to identify pre-seizure biomarkers using wavelet analysis and statistical testing
• Collaborated with a team of neurologists to improve clinical decision-making for epilepsy treatments

• Tech: Python • NumPy • MEG/EEG Analysis Tools
• Link: UCSF
• Date: Fall 2024

Data Science Consultant

Universal Health Services

• Developed time-series financial forecasting models using historical data to predict monthly financial changes
• Trained ARIMA and LSTM models to forecast financial metrics three months in advance, improving forecast accuracy 17% over the existing baseline
• Performed exploratory data analysis, handled missing values, and optimized model performance
• Presented findings to VPs and CFO to inform strategic financial planning

• Tech: Python • scikit-learn • TensorFlow
• Link: Universal Health Services
• Date: Summer 2024

Data Science Researcher

UC Berkeley College of Letters and Science

• Created a logistic regression model with 94.4% accuracy for predicting alumni donation likelihood
• Optimized an LSTM model achieving 86.4% accuracy for donation amount forecasting
• Authored a paper on predictive modeling in alumni donations, presented at Berkeley's Data Science Symposium
• Tech: Python • TensorFlow • Tableau
• Link: Berkeley Letters & Science
• Date: Dec 2023 – May 2024
Close

DeCal Director

UC Berkeley Data Science Society

• Facilitate Data 198: Intro to Real World Data Science DeCal at UC Berkeley leading 11 TAs/Tutors and 50+ students
• Designed and led interactive sessions on data analysis, visualization, and real-world applications
• Mentored students on hands-on projects using Python, Pandas, and Jupyter Notebook
• Previously served as TA (Fall 2024) and tutor (Spring 2024)

• Tech: Python • Pandas • Jupyter Notebook
• Link: DSS DeCal
• Date: Spring 2024 - Present

Machine Learning Researcher

Berkeley AI Research (BAIR), Speech Lab

• Co-authored the Speech World Model, accepted to ICLR 2026, introducing a causal-graph framework that unifies speech understanding and reasoning through interpretable modular inference
• Developed a semi-supervised causal inference pipeline that improved reasoning accuracy by 49% while reducing compute cost to 20 GPU hours on large-scale speech datasets
• Fine-tuned LLaMA 3.1-8B and Qwen2-Audio on graph-guided reasoning tasks in PyTorch, surpassing open-source baselines

• Tech: PyTorch • LLaMA 3.1 • Qwen2-Audio • Python
• Links: Paper (ICLR 2026), BAIR
• Date: Jul. – Dec. 2025

President

Data Science Society at Berkeley

• Led Berkeley's premier data science and ML organization, overseeing 100+ members across education and student-industry initiatives
• Organized Berkeley's 7th Annual Datathon for 500+ participants, coordinating sponsors, speakers, logistics, and competition execution
• Guided the organization's strategy across its education, consulting, and community programs

• Focus: Leadership • Data Science • Community
• Link: Data Science Society
• Date: Dec. 2025 – May 2026

Software Development Engineer Intern

Amazon

• Built an interactive AI coding assistant, selected for production deployment, that converts financial SOP steps into review-ready automation units and cuts development time from a week to hours
• Designed and deployed a separate headless agent on EC2, with DynamoDB-backed workflow state, to autonomously select unautomated SOP items, generate and validate code, and publish code reviews
• Built MCP tooling over internal code, documentation, and API metadata to ground both agents, and an independent validation agent that checks citations, API validity, and 80%+ test coverage
• Benchmarked against a baseline agent across 15 FinOps services, improving API-operation accuracy from 69% to 100% with 34x more source-code references

• Tech: EC2 • DynamoDB • MCP • Multi-Agent Systems • Python
• Link: Amazon
• Date: Summer 2026

Teaching Assistant

University of California, Berkeley

• Teach 300+ students in Data 102: Data, Inference, and Decisions, UC Berkeley's upper-division course on decision-making under uncertainty
• Host discussions and labs for 50+ students, covering optimization, decision theory, and machine learning
• Build privacy-preserving tutoring and calendar workflows for 300+ students through Git and GitHub Pages

• Focus: Teaching • Inference • Decision Making
• Link: Data 102
• Date: Aug. 2026 – Present

Started Undergraduate Studies

UC Berkeley

• Enrolled in the B.A. program for Computer Science and Data Science
• Began exploring interests in AI, neuroscience, and data science through coursework
• Joined student organizations and clubs to engage with Berkeley's vibrant academic community

• Link: UC Berkeley
• Date: Fall 2023

Computational Neuroscience Researcher

UCSF

• Conducted research at UCSF to optimize seizure detection models for personalized clinical care
• Analyzed MEG/EEG data to identify pre-seizure biomarkers using wavelet analysis and statistical testing
• Collaborated with a team of neurologists to improve clinical decision-making for epilepsy treatments

• Tech: Matlab • Python • MEG/EEG Analysis Tools
• Link: UCSF
• Date: Fall 2024

Computational Neuroscience Researcher

UCSF

• Conducted research at UCSF to optimize seizure detection models for personalized clinical care
• Analyzed MEG/EEG data to identify pre-seizure biomarkers using wavelet analysis and statistical testing
• Collaborated with a team of neurologists to improve clinical decision-making for epilepsy treatments

• Tech: Matlab • Python • MEG/EEG Analysis Tools
• Link: UCSF
• Date: Fall 2024

Data Science Consultant

Universal Health Services

• Developed financial forecasting models using historical data to predict monthly financial changes
• Trained ARIMA and LSTM models to forecast financial metrics three months in advance, improving forecast accuracy 17% over the existing baseline
• Performed exploratory data analysis, handled missing values, and optimized model performance
• Presented findings to VPs and CFO to inform strategic financial planning

• Tech: Python • scikit-learn • TensorFlow
• Link: Universal Health Services
• Date: Summer 2024

Data Science Researcher

UC Berkeley College of Letters and Science

• Created a logistic regression model with 94.4% accuracy for predicting alumni donation likelihood
• Optimized an LSTM model achieving 86.4% accuracy for donation amount forecasting
• Authored a paper on predictive modeling in alumni donations, presented at Berkeley's Data Science Symposium

• Tech: Python • TensorFlow • Tableau
• Link: Berkeley Letters & Science
• Date: Spring 2024

Data 198 Director

UC Berkeley Data Science Society

• Facilitate Data 198: Intro to Real World Data Science DeCal at UC Berkeley leading 11 TAs/Tutors and 50+ students
• Designed and led interactive sessions on data analysis, visualization, and real-world applications
• Mentored students on hands-on projects using Python, Pandas, and Jupyter Notebook
• Previously served as tutor and TA

• Tech: Python • Pandas • Jupyter Notebook
• Link: DSS Decal
• Date: Spring 2024

Full-Stack Developer

UC Berkeley Chinese Student Association

• Developed and deployed a Django web app for member management and event registration
• Implemented Google OAuth 2.0 for secure login and authentication
• Integrated MySQL database for efficient data storage and retrieval

• Tech: Django • MySQL • JavaScript
• Link: CSA
• Date: Fall 2023

Started Undergraduate Studies

UC Berkeley

• Enrolled in the B.A. program for Computer Science and Data Science
• Began exploring interests in AI, neuroscience, and data science through coursework
• Joined student organizations and clubs to engage with Berkeley's vibrant academic community

• Link: UC Berkeley
• Date: Fall 2023