Owen Sun Zhang

Computer engineering student at Boston University, class of 2027, building across machine learning, cloud infrastructure, and hardware.

owenz@bu.edu · linkedin.com/in/owen-sun-zhang · github.com/Pine532 · Résumé (PDF) · Boston, MA

Selected work

• From backtesting to live predictions. FreeWheel · Revenue Science, Machine Learning Engineering Intern, June 2026 – August 2026. Built a temporal backtesting framework to evaluate production cost-per-acquisition (CPA) models, then calibrated a simpler model that outperformed the production neural network offline, and validated it in live canary auctions. 97% lower mean conversion-ratio error (6.07 to 0.18 with per-event calibration); 2.88× top-20% lift (vs. 2.72× for the production neural network, offline); 0.745 ROC-AUC (vs. 0.733, offline across 9 conversion events); ~$300K projected monthly savings (an estimated 5% reduction in cost per acquisition at constant conversion volume).
• An agent that repairs broken builds. Microsoft Industry Project, Autonomous LLM build-repair agent · in progress, September 2026 – Present. Developing an autonomous LLM agent that diagnoses and fixes compiler, linker, dependency, configuration, and packaging failures across x86_64, ARM/aarch64, and RISC-V.
• Ad serving on Kubernetes. Comcast, Site Reliability Engineering / DevOps Intern, June 2025 – August 2025. Migrated the Sprinkler ad-serving platform from EC2 Auto Scaling Groups to Kubernetes on Amazon EKS, automated its deployment and scaling, and built its SLO dashboards. EC2 to EKS platform migration (across multiple environments); Zero manual scaling steps (Helm, ArgoCD, and KEDA autoscaling); 99.99% availability target (supported by multi-AZ redundancy).
• Precharge-capable PCB design. Terrier Motorsport · BU Formula SAE Electric, PCB design and low-voltage review, January 2024 – Present. Worked on the design and assembly of precharge-capable PCBs, and reviewed low-voltage designs for power transfer and operator safety.
• Thruster design and review. Boston University Rocket Propulsion Group, Thruster design, January 2024 – Present. Contributed to thruster design on a six-person team, and presented the preliminary and critical design reviews that secured approval for construction and testing.

Education

Boston University, College of Engineering. B.S. in Computer Engineering. Expected May 2027. GPA 3.84.

Selected coursework: Algorithms, Learning From Data, Discrete Stochastic Processes, Computer Organization. Activities: Terrier Motorsport, Rocket Propulsion Group, Chi Phi, Study abroad in London.

Course project: Cancer drug-sensitivity models. Team project for Learning From Data: regression baselines, gradient boosting, and a neural network compared on GDSC and secondary-screen drug-response data, with explicit leakage controls.

Certifications

• AI & Data Science Program. MIT Institute for Data, Systems, and Society (IDSS), August 2026.
• AWS Certified Solutions Architect – Associate. Amazon Web Services, December 2023 – December 2026.
• Google Cybersecurity Professional Certificate. Google · Coursera, October 2023.
• Missouri Seal of Biliteracy, Mandarin. Missouri Department of Elementary and Secondary Education, April 2023.

Technical toolkit

• Languages: Python, C++, SQL, Java, JavaScript, MATLAB.
• ML & data: TensorFlow, TensorFlow Serving, scikit-learn, LightGBM, pandas, NumPy, Snowflake.
• Systems & cloud: Linux, Docker, Kubernetes, AWS (SageMaker, S3, EKS), Terraform.
• Developer tools: Git, Helm, ArgoCD, KEDA, Prometheus, Grafana.
• Spoken languages: English, Mandarin.

Earlier

St. Louis Modern Chinese School, Website Programmer, Team Lead, January 2023 – August 2023. Led a 20-person team through the design and development of the school’s course-registration website, a site serving 1,000+ monthly visitors, and built its back-end registration workflows.

The full site, with its simulated Sun, needs JavaScript.