Evan Mildenberger

Evan Mildenberger

Mentor
Rising Codementor
US$6.00
For every 15 mins
free badge
First 15 mins free for your first session
ABOUT ME
Enthusiastic Full Stack Software Engineer
Enthusiastic Full Stack Software Engineer

I'm currently working in the financial sector using Rust and Terraform, but most of my experience is in full-stack Web development using technologies like Django, Vue and Docker to create single-page and server-side-rendered applications. I've also worked on projects for data science and machine learning, building neural networks from scratch to do things like predict people's movements based on video footage. I'm currently expanding my education to focus on higher-performance computing and embedded environments using the great programming language Rust!

I like projects where I'm accountable for making an impact and can contribute to big-picture decisions. I'm a nerd for math and all obscure tech and science stuff. All my life I've been interested in understanding the universe and the people in it as well as optimizing both human-to-human and human-to-machine communication.

My future endeavours include combining my experience in computer science with my experience in linguistics and intrapersonal communications to create novel solutions for persistent problems. I'm open to opportunities for partnership in entrepreneurship as I prefer agile and creative environments over enterprise ones.

Spanish, English
Tbilisi (+04:00)
Joined August 2019
EXPERTISE
6 years experience
3 years experience
4 years experience
3 years experience
1 year experience
5 years experience
3 years experience

REVIEWS FROM CLIENTS

Evan's profile has been carefully vetted and approved as a Codementor. Connect with Evan now, and leave a review for them once you're done!
SOCIAL PRESENCE
GitHub
tango_trajectory
A completed data science project for a research team. The goal was to forecast the future positions of a tango dancer given the past positions of the other dancer.
HTML
0
0
neural_network
This is a neural network built from scratch using only Numpy.
Python
0
0
EMPLOYMENTS
Principle developer
DCP Global
2023-02-01-Present
- Planned, designed, implemented and deploy a Python project for a recruitment platform with over 3,000 users - Reduced business costs fo...
- Planned, designed, implemented and deploy a Python project for a recruitment platform with over 3,000 users - Reduced business costs for infrastructure by 2500% monthly by migrating compute and database to cheaper alternative
Django
PostgreSQL
Docker
View more
Django
PostgreSQL
Docker
Vue.js 3
View more
Senior developer
CIE
2022-06-01-Present
Designing and implementing Python code for an e-learning platform with over 10,000 users
Designing and implementing Python code for an e-learning platform with over 10,000 users
Django
Celery
Docker
View more
Django
Celery
Docker
Vue.js 3
View more
PROJECTS
Neural Network from ScratchView Project
2020
This was a proof of concept for using a machine learning neural network to classify the MNIST dataset of handwritten digits. This impleme...
This was a proof of concept for using a machine learning neural network to classify the MNIST dataset of handwritten digits. This implementation doesn't rely on high level libraries such as TensorFlow, PyTorch, Keras, etc but uses only Numpy for optimized numeric matrix and vector operations. It implements vanilla backpropagation using first principles of matrix calculus with just enough parameterization to specify the architecture for the width and depth of the dense feedfoward layers. It does not utilize convolutional layers, although that would improve performance.
Python
NumPy
View more
Python
NumPy
View more
Tango Trajectory -- Dancer Movement Forecasting using a LSTM Neural NetworkView Project
2020
This was a contribution to a research paper about the use of LSTM neural networks to forecast future movements of dancers (Tango for this...
This was a contribution to a research paper about the use of LSTM neural networks to forecast future movements of dancers (Tango for this case) using video footage of their past movements taken from above. The video data had to be processed to extract the 2D position time-series data for the two participants. Then this data had organized into training and testing partitions that respected the time-series nature but would also allow sufficient cross validation and training-time validation. Common techniques for finding hyperparaters were used via the mentioned libraries, and the results of various results and the description of the problem were recorded in a Jupyter notebook.
Python
NumPy
Pandas
View more
Python
NumPy
Pandas
Jupyter
Keras
Scikit-learn
View more