PDF FullPDF 1 PageWord FullWord 1 Page


Exactrics, Inc. • Founder & CTO • Jul 2018 — Dec 2019

Exactrics lets you scan clothes with your phone, then search, filter, and compare by fit while shopping online

  • Built a pipeline for converting sets of smartphone photos into millimeter-precision clothing scans
  • Built custom fit-search, look-search, and fit-comparison engines for use with our scan data
  • Built and maintained our backend server and cloud deployment infrastructure
  • Collaborated on the frontend, including development of client SDKs and custom data visualizations

Palantir Technologies • Member of the Technical Staff • Oct 2015 — Feb 2018

  • Adapted modern math, stats, and CS techniques to build solutions for Palantir’s most critical customers and product teams. Selected projects include:

    • A personalized document feed used by intelligence analysts to surface urgent new reports
    • A viewership model used by a TV network to simulate and optimize promotional campaigns
    • A customer churn model used by a major payment processor to prioritize CRM efforts

Schlumberger • Software Engineering Intern • Jun 2015 — Oct 2015

  • Developed algorithms to leverage autocorrelations in time-series data for predictive maintenance
  • Created a time-series-oriented data visualization library for internal use
  • Configured and documented a Spark cluster for use by Schlumberger’s various data analytics teams


Selected Projects

Clothing Fit Scan Pipeline (Exactrics)

  • Reconstruct multiple surface views with traditional photogrammetry techniques
  • Use output masks from CV models to stitch reconstructions into item surface:
  • Use custom U-Net to segment clothing from background in each view
  • Use a Siamese CNN to identify surface correspondences
  • Compute geodesic-based vectors for fit-based search and recommendation

Intelligence Document Recommendation Engine (Palantir)

  • Vectorized documents using various domain-specific features:

    • Custom word vectors built from domain corpus
    • Sparse named entity features (using domain-specific NER refinements)
    • Geographic features from a hidden layer of an auxiliary network with lat/lon inputs
  • Combined document vectors with user interaction data to build user vectors
  • Recommended documents by predicting interaction for (user, document) pairs


Machine Learning and Data Science

  • Have deployed models in numerous domains:

    • TV advertising, drug discovery, fraud detection, remote sensing, O&G predictive maintenance, etc.
  • Experienced with various forms of unstructured data: images, audio, time-series, text, etc.
  • Experienced writing data processing pipelines of various complexities and sizes (up to terabyte-scale)
  • Experienced with out-of-core techniques for “big” data, and optimal inference techniques for “small” data

Computer Vision and Augmented Reality

  • Have used CNNs, U-Nets, and custom neural nets to automate various image-processing tasks
  • Extensive experience using scikit-image and OpenCV for 2D image pre-processing and manipulation
  • Have used OpenCV, Open3D, AliceVision, CGAL, etc., to solve various 3D mesh geometry problems
  • Experienced working with raw RGB-D data from Intel RealSense cameras and dual-camera phones
  • Understand the capabilities and limitations of the high-level mobile AR toolkits (ARKit and ARCore)

High-Performance Python

  • Deeply familiar with scientific python, including numpy, scipy, matplotlib, pandas, scikit-learn, etc.
  • Experienced using Cython and C++/PyBind11 to write various high-performance extensions
  • Experienced orchestrating streaming processing jobs on non-server nodes with celery
  • Experienced using Apache Spark to perform large-scale analytics jobs on a dedicated cluster

Web APIs, Frontend, and DevOps

  • Experienced designing, developing, and deploying APIs for web and mobile front-ends
  • Extensive experience with python’s ASGI server ecosystem, including contributions to pydantic, fastapi, and starlette
  • Experienced with frontend frameworks: React and Vue on the web; UIKit and SwiftUI on iOS
  • Familiar with typical infrastructure requirements and security concerns for public-facing web services
  • Experienced using Docker to develop and deploy containerized web services to multiple cloud providers

Agile Development

  • At Palantir, used scrum to establish a reputation of consistently delivering high-quality results on time
  • Understand the importance of communication; high-frequency iteration with stakeholders is a top priority
  • Methodical about formulating, testing, and confirming/rejecting business hypotheses

Articles and Publications • +1 (650) 842 0413
928 Mouton Cir, East Palo Alto, CA 94303

Created with pandoc_resume