Hi, I am Carlos

BNU Business School student

Experienced Software Engineer & Product Specialist

Case Studies Experience
Dance
Beijing
Seattle
Study
Scroll

What I Do

01

Data-Driven Decisions

Transforming complex data into clear, measurable financial improvements.

02

AI for Business

Leading AI projects from conception to production deployment with measurable impact.

03

Cross-Functional

Aligning specialized departments on project milestones and long-term delivery.

04

Global Perspective

Communicating across cultures for seamless collaboration and product development.

Trusted experience from

· ·

About Me

I am a Master’s researcher at Beijing Normal University Business School and a former AWS Software Engineer.
At Amazon, I led AI and Data Science initiatives in fraud detection and consumer behavior, where I discovered my true edge: bridging the gap between technical engineering and business strategy.
I am currently looking for Product-related positions in the technology or finance sectors.

Education

Beijing Normal University

Master in World Economics

Beijing, China GPA: 3.8/4.0
Sep 2025 - Jun 2027

Duke Kunshan University

B.Sc. in Data Science

Kunshan, China GPA: 3.6/4.0
Sep 2019 - May 2023

Duke University

B.Sc. in Interdisciplinary Studies

Durham, NC, USA GPA: 3.6/4.0
Sep 2019 - May 2023

Professional Experience

Amazon Web Services

Software Development Engineer

Seattle, USA | Jul 2023 - Sep 2025
  • Used Python and SQL to detect anomalies across large-scale usage datasets, uncovering a ∼$500K billing discrepancy. Validated the impact through quantitative analysis and led a cross-functional remediation effort with Finance and Engineering.
  • Built and back-tested time-series forecasting models in Python; applied feature engineering on telemetry data to improve cloud capacity monetization, reducing annual CapEx by ∼$1.5M.
  • Led cross-team execution for the launch of AWS Lambda Managed Instances — became the go-to subject matter expert, resolved blocking dependencies, and authored internal best practices that were adopted across the organization.

Duke Kunshan University

Data Science Research Assistant

Kunshan, PRC | Sep 2022 - May 2023
  • Built and open-sourced a vision-language dataset of over 10 million text-image pairs to support model training and community research.
  • Optimized a text-image alignment model by rethinking the training pipeline — cutting dataset size by 50%, training time from 12 to 5 days, and GPU usage by over 90% with less than 10% performance loss.
  • Applied Deep Learning and Reinforcement Learning to build an automated asset management strategy for DeFi token markets.
  • Validated the strategy through backtesting, achieving over 5% return on investment across historical market data.

Duke University

Research Assistant

Durham, NC, USA | Jun 2021 - Aug 2022
  • Designed and analyzed survey data to identify barriers to social program enrollment for Hispanic immigrants.
  • Created audience-friendly data visualizations and presented findings at a community-focused conference.

Product Case Studies

Selected examples of how I identify business problems, drive cross-functional execution, and deliver measurable outcomes.

AWS Data Analysis Finance

Billing Anomaly Detection & Revenue Recovery

~$500K Recovered annually

Problem

AWS had a systemic billing discrepancy pattern across large-scale usage datasets with no systematic monitoring, leading to significant revenue leakage.

My Role

Led data analysis and remediation. Owned quantitative validation and cross-functional coordination with Finance and Engineering teams.

Process

  • Used SQL + Python to analyze telemetry and identify anomaly patterns across millions of records.
  • Quantified financial impact to secure leadership buy-in from Finance stakeholders.
  • Coordinated remediation across Engineering (root cause fix) and Finance (account reconciliation).

Outcome

Recovered ~$500K in billing discrepancies and established ongoing monitoring practices to prevent future leakage.

Python SQL Data Analysis Cross-Functional Leadership Financial Acumen
AWS Machine Learning Cost Optimization

ML-Driven Cloud Capacity Optimization

~$1.5M Annual CapEx saved

Problem

Cloud capacity was either over-provisioned (wasting CapEx) or under-provisioned (risking outages). Annual waste was estimated in the millions with no data-driven planning.

My Role

Built and validated forecasting models. Owned the feature engineering pipeline and back-testing methodology before production deployment.

Process

  • Applied feature engineering on telemetry data to improve model input quality and predictive accuracy.
  • Built and back-tested time-series forecasting models in Python against historical data.
  • Validated model accuracy and business impact before recommending production deployment.

Outcome

Reduced annual CapEx by ~$1.5M through optimized capacity planning, with models maintaining high accuracy in production.

Python Machine Learning Time-Series Forecasting Feature Engineering Cost Optimization
AWS Infrastructure Cross-Functional

Core Infrastructure Ownership in Cross-Functional Launch

Org-Wide Best practices adopted

Context

AWS Lambda Managed Instances was a major product launch requiring coordination across multiple service teams with complex infrastructure dependencies.

My Role

Owned a critical piece of the core infrastructure. Served as the subject matter expert (SME) for my domain across collaborating teams.

Process

  • Resolved blocking dependencies across teams to keep the launch timeline on track.
  • Authored internal best practices that were adopted organization-wide beyond my immediate team.
  • Served as the go-to expert for cross-team integration questions and technical alignment.

Outcome

Successful launch of a major AWS feature. My best practices became organizational standards, reducing onboarding time for future cross-team projects.

Infrastructure Ownership Cross-Functional Collaboration SME & Knowledge Sharing Dependency Resolution Influence Without Authority
Currently Building

Mandarin Learning Platform

A personal project exploring AI-powered language learning — applying product thinking from problem identification to MVP development.

Research & Publications

IEEE Conference 2022

A Data Science Pipeline for Algorithmic Trading: A Comparative Study of Applications for Finance and Cryptoeconomics

L. Zhang, T. Wu, S. Lahrichi, C.G. Salas-Flores, J. Li

2022 IEEE International Conference on Blockchain

View Publication ↗
ICML Workshop 2023

Making Text-Image Connection Formal and Practical

L. Zhang, C.G. Salas-Flores, D. Zou

Neural Compression Workshop at ICML 2023

View Publication ↗

Certifications

HSK 4 Chinese Proficiency (253/300)

Confucius Institute

Issued July 2026

TOEFL iBT (118/120)

ETS

Issued June 2026

Advanced Programming in C++

Peking University, Coursera

Issued 2020

IBM AI Engineering

IBM, Coursera

Issued 2020

Honors & Awards

Full Merit Scholarship: Silk-Road Muduo

Beijing Normal University | 2025

Awarded for academic excellence and research potential.

Duke Hackathon 2nd Place

Major League Hacking (MLH) | 2022

Built "Blinder Resume," an AI-driven hiring tool designed to eliminate unconscious bias.

National Finalist: Alibaba GET Global Challenge

Alibaba Business School | 2020

Developed a mental health platform; outperformed 250+ teams to reach the China National Finals.

Dean's List

Duke Kunshan University | 2019, 2020

Recognized for maintaining a GPA within the top 10% of the cohort.

Full Tuition Scholarship

Duke Kunshan University | 2019 – 2023

Four-year merit-based scholarship covering 100% of tuition expenses.

Skills & Tools

Product Management

Product Strategy Roadmapping A/B Testing User Research PRD Writing Stakeholder Management Metrics & Analytics

Data & Engineering

Python SQL Data Analysis Machine Learning Deep Learning Time-Series Forecasting AWS (Lambda, S3, etc.)

Tools

Jira Confluence Git Excel

Languages

English (Native) Spanish (Native) Mandarin Chinese (Professional)

Soft Skills

Cross-Functional Leadership Bilingual Communication Global Team Collaboration

Beyond the Code

The person behind the product — how diverse experiences shape my approach to building and leading.

Street Dance

Street Dance

3+ years of competitive breakdancing taught me discipline, creativity under pressure, and performing in high-stakes environments.

Traveling

Global Explorer

Living across Asia and the Americas shaped my ability to navigate cultural differences — a skill I apply daily in cross-functional product teams.

Languages

Multilingual

Learning Mandarin to professional fluency (HSK 4) reflects my commitment to deep cultural immersion, not just surface-level adaptation.

Get in touch

Looking for PM opportunities in Tech & Finance.

WeChat: YUSEICARLOS

Notes