Eric Sinzinger, Developer in Los Angeles, CA, United States
Eric is available for hire
Hire Eric

Eric Sinzinger

Verified Expert  in Engineering

Computer Vision Developer

Location
Los Angeles, CA, United States
Toptal Member Since
December 11, 2016

Eric has been a computer science professor, a researcher for the US government, and a software engineer for startup companies. He has excellent problem-solving and communication skills, believes in making the code base as small and simple as possible, and has worked across several varieties of back-end stacks. Eric specializes in data science, machine learning, and code optimization.

Portfolio

Xometry
ETL, Refactoring, Microservices, Python 3, Amazon Web Services (AWS), Scrum...
United States Government
SQL, Python, C++, Linux
Texas Tech University
Simultaneous Localization & Mapping (SLAM), 3D Reconstruction, Pattern Matching...

Experience

Availability

Part-time

Preferred Environment

Cloud, Amazon Web Services (AWS), GitHub, C++, Python, MacOS, Windows, Linux

The most amazing...

...project I've developed is a robot that can determine its location and identify a small set of objects.

Work Experience

Director of Computational Geometry | Senior Computational Scientist

2017 - 2020
Xometry
  • Updated regression testing to reduce API failure rate on deployments (2.2% to 0.18%).
  • Evaluated data quality, identified core weaknesses, and replaced in-house mesh repair software with third-party solutions (reduced error rate from 1 in 100 to 1 in 5,000).
  • Created a pipeline to automatically label data from the operations team to reduce false positives and negatives in automatic analysis of manufacturing.
  • Tripled the number of geometric features, while maintaining consistent execution time (together with data quality, reduced the uncertainty in core statistical models by 15%).
  • Improved team performance by removing constraints, streamlining testing, and facilitating code reviews to enable continuous deployment (14 days of integration and testing to 1 hour).
  • Isolated core code into a micro-service to identify performance issues and improve (increased speed by a factor of ten). Resulted in drastically improved customer experience with customers making 70% more modifications to designs.
  • Developed novel real-time algorithms including surface reachability, minimal point distance, and sheet unbending for sheet surface area.
Technologies: ETL, Refactoring, Microservices, Python 3, Amazon Web Services (AWS), Scrum, Kanban, Agile, Machine Learning, NumPy, Python

Software Engineer

2009 - 2017
United States Government
  • Created an automated tool to detect and extract text from videos in real-time using C++, OpenCV, support vector machines, and a novel representation of the text space.
  • Built an organization-wide Linux file-system that allows all data to be tracked, properly categorized, and legally compliant—meeting government requirements for the data storage of critical information.
  • Developed an automated system for the machine language translation of small phrases that filled a need not satisfied by either simple word lookup or more advanced translation of full sentences.
  • Wrote Python scripts to support the automated processing of complex tasks for non-technical customers.
  • Served as the team lead for a vital tiger team that required understanding several different skill sets.
  • Received several team awards, two invention awards, and a patent.
Technologies: SQL, Python, C++, Linux

Associate Professor of Computer Science

1999 - 2009
Texas Tech University
  • Worked as the founding director of the Data Representation and Intelligent Learning Lab and supervised twenty masters students and two PhD students.
  • Taught courses on programming, analysis of algorithms, operating systems, numerical methods, computer graphics, computer visualization, pattern recognition, and computer vision.
  • Developed a novel method for estimating livestock body mass from laser imaging that resulted in licensing and commercialization of the technology.
  • Created a novel method for improving the shape reconstruction from multiple images using an improved data analysis of the neighborhoods around critical points inside of an image. This technology was also licensed for commercialization.
  • Built a computer vision library to support complex mathematical operations on geometric objects that wrapped fast Fortran code underneath a clean object-oriented library in C/C++.
Technologies: Simultaneous Localization & Mapping (SLAM), 3D Reconstruction, Pattern Matching, Scientific Computing, Robotics, Computer Graphics, Computer Vision, Python, C, C++

Research Engineer

1998 - 1999
Summus, Ltd
  • Worked as a developer on a video compression team that was designated as the leader in the video-compression quality in bandwidth limited environments.
  • Invented a new method that improved the accuracy of underwater mine detection by 95% by detecting surface slope of ocean waves and removing the effects of light refraction. This resulted in a technology that received the award of 1999 Advanced Imaging Solution of Year.
  • Wrote real-time, cinema-quality, video-compression code for applications with high bandwidth.
  • Optimized the code by detecting code bottlenecks and replacing expensive functions with numerical approximations implemented in x86 Assembly. This resulted in a 20-time speed-up of the specific functions, and a 3-time speed-up of the overall compressor.
Technologies: Wavelets, Assembly Language, C, C++

Flatland

Created a system for the archaeological records of 3D scanned historical objects. Although the data was 3D, the archaeologists were more comfortable with 2D printings. Geo-rectified the 3D data to a cardinal 2D coordinate system and unwrapped 360 degree views into a flat representation allowing archaeological preservation of national park landmarks.

Object Recognition for Crew Exploration Vehicle

While working as a visiting scientist at NASA, developed a system for real-time recognition of complex 3D shapes. Relied upon the crew exploration vehicle only having a small set of objects predetermined prior to launch, then developed shape classifiers for each object based upon surface characteristics of each shape.

Languages

Python 3, C, C++, Python, JavaScript, Fortran, Assembly, SQL, Assembly Language

Libraries/APIs

OpenGL, OpenCV, Pandas, NumPy, Dask, D3.js

Tools

LabVIEW, Git, MATLAB, GitHub, Bamboo

Paradigms

Refactoring, ETL, Wavelets, Continuous Integration (CI), Continuous Delivery (CD), Scrum, Kanban, Agile, Data Science, Management, DevOps, Microservices, Microservices Architecture

Platforms

Linux, Amazon EC2, iOS, Docker, Amazon Web Services (AWS), Windows, MacOS

Storage

SQLite, MySQL, Amazon S3 (AWS S3), MongoDB

Other

Artificial Intelligence (AI), Neural Networks, Deep Neural Networks, Convolutional Neural Networks (CNN), Machine Learning Automation, Machine Learning Operations (MLOps), Data Engineering, 3D Reconstruction, Image Processing, Cloud, Simultaneous Localization & Mapping (SLAM), Numerical Methods, Scientific Computing, Computer Vision, Computer Graphics, Machine Learning, Deep Learning, Back-end, Natural Language Processing (NLP), Robotics, Graphics Processing Unit (GPU), GPU Computing, Full-stack, Architecture, GPT, Generative Pre-trained Transformers (GPT), Pattern Matching, Sequence Models, Natural Language Understanding (NLU), Recurrent Neural Networks (RNNs)

Frameworks

Spark, Apache Spark

1996 - 1999

Ph.D. in Computer Science

University of South Carolina - Columbia, SC, USA

1996 - 1998

M.S. in Computer Science

University of South Carolina - Columbia, SC, USA

1994 - 1996

M.S. in Mathematics

University of South Carolina - Columbia, SC, USA

1989 - 1993

B.A. in History

University of Texas at Austin - Austin, TX, USA

1989 - 1993

B.S. in Mathematics

University of Texas at Austin - Austin, TX, USA

JULY 2020 - PRESENT

Deep Learning

Coursera

Collaboration That Works

How to Work with Toptal

Toptal matches you directly with global industry experts from our network in hours—not weeks or months.

1

Share your needs

Discuss your requirements and refine your scope in a call with a Toptal domain expert.
2

Choose your talent

Get a short list of expertly matched talent within 24 hours to review, interview, and choose from.
3

Start your risk-free talent trial

Work with your chosen talent on a trial basis for up to two weeks. Pay only if you decide to hire them.

Top talent is in high demand.

Start hiring