Bay Area · Data Science · AI / ML
also known as Jennica
I build machine learning systems, predictive models, and data-driven tools that work in the real world — from clinical AI to workforce intelligence to on-device ML.
Aye Nyein Kyaw
(Jennica)
Bay Area, CA
I'm a Data Scientist based in the Bay Area with hands-on experience building end-to-end ML systems — from raw data pipelines to deployed products. My work spans computer vision, NLP, predictive modeling, and interactive dashboards across real-world domains.
I've built a clinical AI tool achieving 96.7% pneumonia sensitivity deployed on Streamlit, an early-warning layoff forecasting system combining 4 economic data sources into a live dashboard, an on-device sign-language translator built at a Google × Qualcomm hackathon, and a financial NLP system using FinBERT + XGBoost for market signal detection.
I don't just build models — I make sure they're explainable, responsible, and production-ready. Every project ships with documentation, visualizations, and real users in mind.
"Turning messy, uncertain things into something organized and meaningful — that's what data science means to me."
96.7% sensitivity. Real X-rays. Real patients. A clinical AI tool deployed for a children's hospital in Myanmar.
Fine-tuned ResNet18 on 5,200+ pediatric X-rays. Generates AI clinical reports via a RAG pipeline pulling live WHO/CDC guidelines. Ships with Grad-CAM heatmaps and a 70% confidence gate for responsible AI.
4 data sources. 1 live dashboard. Predicts California layoff risk before it happens.
Fuses WARN records, FRED economic indicators, GDELT news sentiment, and tech layoff data into a unified monthly forecasting pipeline. Multi-page Dash app with EDA, methodology, and live risk predictions.
No internet. No server. No barriers. Real-time sign language translation built for offline accessibility.
On-device sign language translator running fully offline on Android via LiteRT/TFLite and Qualcomm AI Hub NPU. Built in 48 hours for privacy-first, real-time communication support.
FinBERT + XGBoost + SARIMAX. Three models, one mission — read the market before it moves.
Fine-tuned FinBERT with XGBoost to classify financial market signals from news headlines. Combined NLP features with time-series modeling for predictive accuracy and responsible signal interpretation.
I'm actively seeking full-time roles in data science, ML engineering, and analytics. I'd love to hear from you.