Po Hsuan Ho

Professional machine learning and data scientist with strong business acumen and client-facing sales experience. Proficient in Python, SQL, SAS, and ML frameworks, with 5+ years in AWS environments. Former opera singer, offering exceptional stage presence, confident public speaking, and the ability to connect with diverse audiences. Adept at strategic market analysis, building rapport, assessing customer needs, designing technology solutions, and driving sales.

Business & Art

Opera Singer/ Music studio Owner Sep 2012 – May 2015

• Performed leading roles in operas including Chérubin (Title role), Beth (Little women), Cherubino (Le nozze di Figaro), Despina (Così fan tutte) in 600-1000 seats concert halls in Minneapolis, Baltimore, Toledo, Maine, and Arcidosso in Italy. 
• Provided vocal performance and piano instruction to over 50 students, aged 4 to 68, through individual lessons and group classes.

Technical Sales Representative, King Wang Tool Jun 2015 – Sep 2018

• Managed B2B sales by identifying and gathering requirements from 20+ major stakeholders, fostering strong relationships, and collaborating with cross-functional business and technical teams to align product designs with client needs.
• Increased account growth by 35% through in-depth market research and alignment of product developments with customer business goals.
• Designed and implemented management dashboards and predictive analytics tools, driving strategic decision-making processes and supporting client objectives through data-driven insights.

Data Scientist, Sparksoft Corp Mar 2019 – May 2024

• Created impactful visualizations and compelling presentations for Marketplace Program Integrity Contract (MPIC) stakeholders, effectively communicating complex data insights to non-technical audiences.
• Conducted advanced data analysis to identify fraud, waste, and abuse in healthcare claims, utilizing SQL and SAS for segmentation and predictive modeling to optimize client solutions.
• Managed multiple concurrent priorities in a fast-paced, dynamic environment by optimizing onboarding and training processes. Developed comprehensive documentation and SOPs, enabling new developers to achieve standard delivery speeds within 2 weeks.
• Collaborated with stakeholders to develop and monitor over 20 critical data indicators and KPIs, creating a composite model that drives ongoing investigations.
• Detected anomalies in healthcare claim data on Amazon Redshift using predictive models, including logistic and linear regression. Enhanced the accuracy of identifying fraudulent agents/brokers by 15%.

• Yelp Review Rating Prediction: Built machine-learning models to analyze Yelp reviews with NLP and achieved > 90% accuracy.
• Stack overflow posts analysis using Apache Spark: Investigated over 24 million XML file posts, analyzing the relationship between post favorite count and scores, user reputation and post types, and factors for user activity over time. Developed ML pipelines to streamline the fitting of various models and predict if a question falls under the common tags.
• Dental cavity detection on radiographs: Built an end-to-end cavity detection model utilizing transfer learning based on the VGG16 model, and deployed the model on a Streamlit-based interactive website, allowing users to upload radiographs for automatic identification and highlighting of affected areas.

Reduced over 20% of costs through instance conversion and predictive usage analysis by cleansing AWS service cost data and designing interactive dashboards for 20+ teams.

M.S. in Data Analytics | Harrisburg University - Dec 2021

B.A. in Music: Opera and Music Education | National Taiwan Normal University - May 2012