Machine learning · Research · Engineering

Chuck Cho

cycho@cycho.orgGitHub / cycho

Summary

Accomplished machine learning engineer with extensive experience in applied ML (computer vision, speech/language/multimodal, and ranking/recommendation). Proven track record of AI/ML research, prototyping, and productionization. Strong ability to manage senior-level relationships and cross-functional teams.

Experience

Meta Facebook, Bellevue, WA (2020 - present)

  • Delivered a next-gen content understanding model for all incoming videos on Facebook/Instagram which is trained on billion-scale data and outperforms the previous production model by >15% mAP.
  • Developed new ML-based approaches to detecting and mitigating scraping activities on Meta platforms with 98% precision.
  • Led Search Engine Results Page (SERP) optimization for Facebook Groups, and drove 3.5M user (DAU) gain.

Axon, Seattle, WA, Dextro, NY, NY (2015 - 2020)

  • Led development of AI-empowered products for the law enforcement domain, including topic classification for body-worn camera footage, speech-to-text service, and acoustic event classification.
  • Managed and coached a team of 7 Ph.D. ML researchers. Established external collaborations with MIT CSAIL and AI ethics scholars.
  • Delivered video classification system including designing video-specific classifier models, model fusion schemes, and evaluation pipeline. Won a video analytics project for LAPD.

SRI International, Princeton, NJ (2009 - 2015)

  • Implemented and delivered Improvised Explosive Device detection systems using EO/IR videos, hyperspectral SWIR sensors, and coded aperture snapshot imagery. A resulting Night Eagle system helped save hundreds of lives.

IBM Semiconductor Research and Development Center, East Fishkill, NY (2005 - 2009)

  • Led and developed Design for Manufacturing (DfM) methodologies and process-induced variability modeling in advanced Silicon technologies from 90nm to 45nm. Authored >20 technical papers and >10 patents.

Remote Sensing and Estimation Group, Research Laboratory of Electronics, MIT, Cambridge, MA (2002 - 2005)

  • As a graduate research assistant, developed an ML-based solution to extract underlying physical parameters from raw weather satellite imagery data.

MIT Information Systems, Cambridge, MA (2001 - 2005)

Working as an Athena consultant. Researched/answered student and faculty questions regarding the use of software on MIT's campus-wide computer network.

Multimedia Processing and Communication Lab., Korea University, Seoul, Korea (1997)

Assisted in the development of the algorithms for a digital image stabilizer as an undergraduate research assistant, funded by Samsung Electronic Corporation.

POSCO, Kwangyang, Korea (1997)

Supported a research team engaged in installing and managing new communication networks for factory automation.

Software

Owner / maintainer of video-caffe at github, a fork of Caffe that can handle video and Nd-convolution/pooling.

Publications

Imaging / Computer Vision

Process variability modeling / design-for-yield (DFY)

Circuit

Device / technology

Remote sensing / detection and estimation

Patents

Invited talks

  • Linking Product and Manufacturing Via a Statistical Framework, presented at Communication Technologies Department Seminar at IBM T.J. Watson Research Center seminar, July 2007.
  • Anomaly Compensation and Cloud Clearing of AIRS Hyperspectral Data (ppt), presented to the meeting of IEEE Geoscience and Remote Sensing Society (GRSS) Boston Section, August 2005.

Skills

  • Formerly served as a webmaster and system administrator of MIT Korean Graduate Students Association from 1999 to 2001.
  • Programming experience in shell, C/C++, Java, Matlab, SAS, XML/HTML/PHP/CSS, SQL, Perl, Tcl/Tk; Proficiency in the Unix system administration.
  • Web design / web server administration since 1995. I designed and maintained several websites including cycho.org, and kncchurch.org.

Membership / Activity

Education