Close

Mark Fajet

Machine Learning Engineer

View Resume Play with an ML model

About Me

My name is Mark Fajet. I'm a machine learning engineer and software developer currently working for Amazon Web Services as a software development engineer 2 with a passion for mathematics and algorithms. I spend my free time improving my software engineering and machine learning skills, playing guitar, and taking care of my dog and blue-tongued skink.


In 2018, I graduated from Florida International University with a Bachelors of Science in Computer Science and a separate Bachelors of Science in Mathematical Sciences with Summa Cum Laude. I enjoy learning and I enjoy helping others learn as well so, after spending over a year in my current role with AWS, I decided to pursue a Master of Science in Computer Science at Florida International University while continuing to work remotely with AWS.


Now, as I come to the end of my graduate program, I am looking for a role I can be passionate about that blends my computer science and mathematics background to solve machine learning problems efficiently and creatively.

Experience

Amazon

Software Development Engineer 2

  • Built recommender system using MLxtend for internal operations portal
  • Integrated search engine for internal operation documents using Amazon Elasticsearch Service, Amazon S3, AWS Lambda, and AWS CDK
  • Defined and collected metrics to be used for data analysis
  • Participated in internal Machine Learning University

Amazon

Software Development Engineer 1

  • Implemented a web application using Flask and React for EFS operators to more easily triage issues, read runbooks, and view real-time metrics with 130 daily users
  • Added multithreading, multiprocessing, and caching to a variety of existing tools which drastically improved performance by up to 94%
  • Improved test coverage and practices with Jest, pytest, and Cypress
  • Mentored interns and new hires during on-boarding and project completion
  • Handled multiple large scale migrations of tools

MathWorks

Engineering Development Group Intern

  • Carried out deep learning benchmarking in MATLAB, Tensorflow, Keras, and MXNet across multiple systems: CPU, single GPU, and multiple GPUs
  • Conducted exploratory data analysis in MATLAB employing linear regression, K-Nearest Neighbors, decision trees, and SVM to diagnose long startup times

Florida International University

Learning Assistant

  • Assisted students with programming projects for courses that focused on assembly language, registers, binary, logic, circuits, caching, paging, multithreading, and GPU programming

Florida International University Honors College

Web Developer

  • Built and managed websites using technologies such as Node, PostgreSQL, React, PHP, Angular, and Firebase

Florida International University Honors College

Peer Instructor

  • Taught Programming for Non-Programmers, a course where students learned basic concepts of programming such as variables, decision making, functions, and Object Oriented Programming (OOP).

Education

Master of Science in Computer Science

August 2019 - December 2020

Florida International University

Bachelor of Science in Mathematics

August 2014 - May 2018

Florida International University

Activities and societies: Statistics Club, American Statistical Association, Math Club at FIU

Bachelor of Science in Computer Science

August 2014 - May 2018

Florida International University

Activities and societies: Honors college, Association for Computing Machinery, numerous hackathons

Certificates

Machine Learning with Python

September 2018

Coursera License BBLX47UKVKDS

This course provides a general overview of machine learning topics such as supervised vs unsupervised learning, model evaluation, and machine learning algorithms using Python and frameworks such as scikit-learn, SciPy, and NumPy

Deep Learning Specialization

March 2019

Coursera License LJS7ZX7L3PCZ

Group of five courses to teach the foundations of Deep Learning, how to build neural networks, and how to lead successful machine learning projects. The courses go into detail about Convolutional networks, RNNs, LSTM, Adam, Dropout, BatchNorm, and Xavier/He initialization. It includes case studies from healthcare, autonomous driving, sign language reading, music generation, and natural language processing. The courses teach not only the theory, but also how it is applied in industry.

Neural Networks and Deep Learning

March 2019

Coursera License EHKAK5VLYEM5

This course teaches the foundations of deep learning, the major technology trends driving Deep Learning, how to build, train and apply fully connected deep neural networks, how to implement efficient (vectorized) neural networks, and describes the key parameters in a neural network's architecture

View More Certificates

Improving Deep Neural Networks: Hyperparameter tuning, Regularization and Optimization

March 2019

Coursera License 246M3G5A64PR

This course focuses on how to get good results from neural networks. It covers best practices, various algorithms to optimize and improve performance, bias/variance tradeoff, and how to tune hyperparameters in a productive manner.

Structuring Machine Learning Projects

March 2019

Coursera License V3DWZU2FGDNX

This course describes how to build a successful machine learning project. The principals and knowledge in this course provide an undesrtanding of how to diagnose errors in a machine learning system, and be able to prioritize the most promising directions for reducing error. It goes over complex ML settings, such as mismatched training/test sets, and comparing to and/or surpassing human-level performance and how to apply end-to-end learning, transfer learning, and multi-task learning

Convolutional Neural Networks

March 2019

Coursera License 97S845LE2Z2A

This course is about how to build convolutional neural networks and apply it to image or video data. It provides practice using convolutional neural networks for visual detection and recognition tasks, and using neural style transfer to generate art.

Sequence Models

March 2019

Coursera License EXZABE7CJ4V6

This course demonstrates how to build models for natural language, audio, and other sequence data. It explained Recurrent Neural Networks (RNNs), and commonly-used variants such as GRUs and LSTMs. It has assignments that apply sequence models to natural language problems, including text synthesis, and to audio applications, including speech recognition and music synthesis.

Mathematics for Machine Learning: PCA

October 2018

Coursera License AGPCKW5US3ZB

This intermediate-level course introduces the mathematical foundations to derive Principal Component Analysis (PCA), a fundamental dimensionality reduction technique. It covers how to use some basic statistics of data sets, such as mean values and variances, compute distances and angles between vectors using inner products and derive orthogonal projections of data onto lower-dimensional subspaces. Using all these tools, the course shows how to then derive PCA as a method that minimizes the average squared reconstruction error between data points and their reconstruction.

Fundamentals of Deep Learning for Computer Vision

September 2019

NVIDIA Deep Learning Institute License cac52ee52d7d42a8982005bdfffc964b

This course covers how to carry out common deep learning workflows such as Image Classification and Object Detection and experiment with data, training parameters, network structure, and other strategies to increase performance and capability.

Fundamentals of Accelerated Computing with CUDA C/C++

October 2019

NVIDIA Deep Learning Institute License 56092573c3a44cdcac3b20a405319a86

This course covers how to use CUDA to drastically improve the performance of CPU-only applications written in C/C++ by taking advantage of GPUs. It introduces the platform as well as how to do efficient memory management.

Fundamentals of Accelerated Computing with CUDA Python

October 2019

NVIDIA Deep Learning Institute License d2b3a8183cea4fb69f7a9a3039505d7a

This course covers how to use CUDA in python applications to gain performance improvements by utilizing the GPU. It covered Numba, PyCUDA, and when to use each as well as additional memory management techinques.

Functional Programming Principles in Scala

June 2019

Coursera License ZMM9LFG7WFM8

This course covers elements of the functional programming style and learn how to apply them usefully in your daily programming tasks using Scala.

Functional Program Design in Scala

July 2019

Coursera License E4ZEKS943S5X

This course taught how to apply the functional programming style in the design of larger applications, covering a variety concepts, from lazy evaluation to structuring your libraries using monads. It included graded project from state space exploration to random testing to discrete circuit simulators..

Parallel Programming

October 2019

Coursera License DNPZJK487EEE

This course taught how to use scala and functional programming to write parallelized algorithms to solve problems efficiently.

Projects

Codenames DQN

  • Created an OpenAI Gym environment of the boardgame Codenames to simulate the game play utilizing GloVe word vectors
  • Trained a Deep Q-network with the goal of creating a model that can play and win the boardgame, consequently learning the multi-modal definitions of English words

Neural Gonna Give You Up

  • Developed encoder-decoder neural network to process MIDI song files to make them sound more like Never Gonna Give You Up by Rick Astley. This project was awarded Best Artificial Intelligence and Machine Learning Hack at UHack 2017
View Project

Rock Paper Scissors

  • Built and trained a deep learning, image recognition model in MATLAB to determine if one is displaying rock, paper, or scissors in real time using transfer learning techniques
No link because it was created at an internal MathWorks hackathon
View More Projects

Web Page Archiving and Content Analysis Tool

  • Created a desktop application that allows users to download large batches of websites into a custom, compressed file format which allows for easy programmatic access and extracts the metadata and main content of downloaded articles using Natural Language Processing. In collaboration with MIT Lincoln Lab
View Project

Culture FitT

  • Web application created as a recruitment analysis tool. It analyzes tweets from users in order to see if they'd be a good fit of a certain culture. I helped create the backend. I used the Twitter API and Node.js for my part of this project. In addition, I used child processes to run data analysis in R. This project was recognized as top 10 projects at MLH Prime 2016
View Project

ROAM

  • An artificial reality game, Game description can be found in GitHub README. For my part, I wrote a Python microservice in order to retrieve and style a map obtained from the Google Static Maps API. In addition, I did some basic math in order to make a function that allows me to return adjacent areas to that original map so that there is a seamless transition from one map to another. This project won 2nd place at MIA Game Jam in 2016
View Project

Night Watch

  • This project uses EMG and temperature sensors in order to detect muscle movement and breathing. The purpose was for people with disorders to be able to have monitoring as they sleep. I made the mobile application that people could use in order to interface with the data collected by the sensors using React Native and MaterialUI. This project won 2nd place and Best Social Good Hack at MangoHacks 2017
View Project

Internet of Trikes

  • The smartest tricycle. Alexa integration, augmented reality, image processing, headlights, speakers. Created at MangoHacks 2018
View Project
-->

Get in Touch

You can message me on LinkedIn