Konstantin Lvovich Tolskiy, Developer in Redmond, WA, United States
Konstantin is available for hire
Hire Konstantin

Konstantin Lvovich Tolskiy

Verified Expert  in Engineering

Bio

Konstantin is a senior programmer with over 20 years of progressive software development and research experience in the fields of 3D graphics, machine learning, graph theory, and so on. Along with having a strong analytical/mathematical background, Konstantin is the author of 12 scientific papers that primarily focus on numerical methods.

Portfolio

Outlier
Large Language Models (LLMs), AI Model Training
Microsoft
C++17, Windows 10, DirectX, Object-oriented Programming (OOP), TypeScript...
Facebook Reality Labs
Python 3, NVIDIA CUDA, C++17, Linux, GPU Computing, Image Processing...

Experience

  • Calculus - 20 years
  • C++ - 20 years
  • Image Processing - 5 years
  • Python 3 - 5 years
  • NVIDIA CUDA - 4 years
  • Maps - 4 years
  • C++17 - 3 years
  • Deep Learning - 2 years

Availability

Full-time

Preferred Environment

Android, Linux, Windows, Game Development, Lua, Finance, Image Classification

The most amazing...

...project I've worked on was a Python script that created and visualized depth maps based on 2D pictures.

Work Experience

Senior Reviewer

2024 - PRESENT
Outlier
  • Trained LLMs, mostly in the field of mathematics and physics.
  • Created challenging prompts that made models fail.
  • Reviewed responses, wrote sample responses, and found ground truth.
Technologies: Large Language Models (LLMs), AI Model Training

Senior Software Engineer

2020 - 2024
Microsoft
  • Decreased Teams rendering power consumption by 40%.
  • Created a subsystem for image dumping, which simplifies the debugging process.
  • Modified the telemetry system for black and frozen frames detection.
Technologies: C++17, Windows 10, DirectX, Object-oriented Programming (OOP), TypeScript, Architecture, Software Architecture, Code Review, Azure, .NET

Software Engineer IV (Contract)

2020 - 2020
Facebook Reality Labs
  • Designed and implemented software for AR/VR research and 3D scene reconstruction.
  • Checked the quality of reconstruction using Ray tracing.
  • Accelerated tests by 400% without losing quality as well as incorporating additional tests.
  • Converted code from Nvidia CUDA GPUs to Intel CPUs.
Technologies: Python 3, NVIDIA CUDA, C++17, Linux, GPU Computing, Image Processing, Artificial Intelligence (AI), Computer Vision, Object-oriented Programming (OOP), Python, Prototyping, Software Architecture

Senior 3D Software Engineer

2017 - 2019
Allign Tegnology
  • Developed new features for orthodontics and the UI for parameter control.
  • Optimized shapes of the features to create the best experience for patients.
  • Improved the performance and quality of 3D mesh generation.
Technologies: Splunk, Multithreading, OpenGL, Windows 10, Microsoft Visual C++, Object-oriented Programming (OOP), Anomaly Detection, Software Architecture, Code Review, Industrials, Image Manipulation

Software Engineer

2016 - 2017
Bitanimate
  • Built depth maps based on 2D pictures using machine leaning.
  • Developed stereo landscape visualizations based on NASA WorldWind and parsed using Open3D building data.
  • Created stereo landscape visualizations based on Google Earth.
Technologies: Linux, Windows 10, TensorFlow, OpenGL, C#, JavaScript, Java, Python, C++, Deep Learning, Artificial Intelligence (AI), Convolutional Neural Networks (CNNs), Object-oriented Programming (OOP), Anomaly Detection, Statistics, WebGL, Architecture, Prototyping, Software Architecture, .NET, AI Model Training

Software Engineer IV

2014 - 2016
Mentor Graphics
  • Designed and implemented an algorithm for finding cutting pairs in linear time.
  • Developed an algorithm for finding separation pairs in linear time.
  • Composed an algorithm that discovers cutting triplets in near linear time.
  • Improved heuristics for graph-coloring algorithms.
Technologies: Graphs, Linux, C++, Object-oriented Programming (OOP), Mathematics, Statistics, Prototyping, Software Architecture, Code Review, Semiconductors

Software Developer

2010 - 2013
ESRI
  • Designed and implemented 2D and 3D visualization systems for GIS.
  • Built optimization structures for a data exchange which resulted in a 1000% acceleration.
  • Implemented the dynamic creation of 3D objects and developed effective methods for their selection.
Technologies: Multithreading, OpenGL, Direct3D, NVIDIA Nsight Systems, VTune, C++, Visual Studio, NVIDIA CUDA, GIS, Object-oriented Programming (OOP), SQL, Code Review

Senior Software Development Engineer

2006 - 2010
AMD
  • Designed and implemented an application that collects information from graphics libraries wrappers and processes and visualizes it, with features for setting graphic state overloads and bottleneck analysis through frame rate changes.
  • Developed shader debuggers. This work included patching shaders and mapping between the Assembler and HLSL codes.
  • Ported data collectors from Windows desktops to embedded platforms (Windows CE and Embedded Linux) and wrote OpenCL samples.
Technologies: C++, DirectX, Lua, Python 3, Prototyping, Code Review, Semiconductors

Software Engineer

2004 - 2006
Bancware, Inc. (subdivision of Sungard, Inc.)
  • Developed, supported, and debugged code and algorithms for the financial simulation (Bancware AML).
  • Implemented Monte Carlo and deterministic simulations for cash flows, market values, and schedules for instruments like options, mortgages, bonds, and CMOs.
  • Performed evolution of various libraries, code reviews, and performance tuning.
Technologies: C++, Monte Carlo

Experience

Creation of Depth Maps Based on 2D Pictures

I composed a Python script that created and visualized depth maps based on 2D pictures. I was the sole contributor. My work included studying of the state-of-art, designing and implementation CNN models on Windows 10 and Linux platforms. I used AWS severs for training.

Video Separation

This is a video segmentation project. The most complicated part is dividing feet and a floor. Segmentation was done in real-time using cameras that provided in-depth information. The software was implemented using C++ 17, CUDA, and more.

Education

1977 - 1983

Master of Science Degree in Physics

Lomonosov Moscow State University - Moscow, Russia

Certifications

JUNE 2018 - PRESENT

Machine Learning

Stanford University | via Coursera

Skills

Libraries/APIs

OpenGL, TensorFlow, OpenCV, DirectX, WebGL

Tools

Microsoft Visual C++, Splunk, GIS, Visual Studio, VTune, NVIDIA Nsight Systems, Direct3D

Languages

C++, C, Python, C++17, Python 3, Java, C#, Lua, SQL, JavaScript, TypeScript

Paradigms

Object-oriented Programming (OOP), Anomaly Detection

Platforms

Windows, Linux, NVIDIA CUDA, Azure, Android, Amazon Web Services (AWS)

Frameworks

.NET

Other

Linear Algebra, Numerical Methods, Calculus, Windows 10, Multithreading, GPU Computing, Mathematics, Prototyping, Code Review, Graphs, Machine Learning, Deep Learning, Maps, Image Processing, Artificial Intelligence (AI), Computer Vision, Convolutional Neural Networks (CNNs), Game Development, Statistics, Architecture, Software Architecture, Semiconductors, AI Model Training, Image Classification, Image Manipulation, Data Processing, Server Infrastructure, Video Encoding, Finance, Monte Carlo, Industrials, Large Language Models (LLMs), Text to Image AI

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