// ml engineer in nyc

I build ML systems that
make decisions at scale.

At Amazon, I build deep learning models that predict what ads you see across every marketplace. Before that, I owned the auction service, migrated ad serving to streaming, and built the experimentation platform. I like going deep on hard problems -- from model training to the infra that serves it.

Resume (PDF) LinkedIn
// experience

Where I've worked

Machine Learning Engineer 2021 — present
Amazon · Sponsored Products, Ads Ranking & Prediction · NYC
  • Build deep learning models for CTR prediction serving Sponsored Products across all Amazon marketplaces. Led co-serve consolidation to unify models across regions, cutting infra cost while improving prediction quality.
  • Migrated the entire ad serving workflow from RESTful to reactive streaming, enabling multi-response delivery that reduced latency and unlocked new ad formats.
  • Owned Mercury, the Tier 1 auction service processing billions of daily ad requests at 99.99% uptime -- the system that decides what sponsored products surface on Amazon search.
  • Designed ranking algorithms that optimize ad allocation under budget and relevance constraints, contributing to $500M in annual revenue growth.
  • Built an experimentation platform from scratch: automated A/B test lifecycle end-to-end, cut launch time from 2 weeks to 2 days, enabled 200+ org-wide experiments.
Full Stack Engineer 2017 — 2021
EdLab, Teachers College Columbia University · NYC
  • Led development of the EdLab website and CMS used by the Teachers College community -- React/Redux frontend, Node.js backend.
  • Worked with researchers to turn ed-tech ideas into real products. Designed APIs, data pipelines, and analytics tools.
  • Introduced automated testing and CI/CD practices to the team.
// skills

What I work with

Languages
Java Python TypeScript JavaScript SQL
ML & Data
Deep Learning CTR Prediction Model Training Model Serving Feature Engineering A/B Testing
Systems & Infrastructure
Distributed Systems Reactive Streaming Real-time Ranking Microservices Low-latency Pipelines
AWS
CDK Lambda DynamoDB Step Functions S3 CloudFront EC2 SQS
Frontend
React Redux Node.js
// education

Where I studied

Stevens Institute of Technology 2014 — 2016
M.S. Computer Science
Dalian University of Technology 2010 — 2014
B.E. Software Engineering