Head Shot
Austin Lasseter, PhD
Pencil Headshot

Software Developer & Machine Learning Engineer

A life-long learner, I love to share with others the latest advancements in automation of AI/ML in the AWS cloud.

My Resume

My Projects

About Me I'm fascinated by the intersection of Software Development and Machine Learning in the cloud!

Cloud-Native AI/ML Software Engineer

Hey! I'm Austin Lasseter and I'm a software engineer and tech instructor located in Northern Virginia. My specialty is building and deploying machine learning models for my clients. I lead remote & onsite courses in data science, artificial intelligence, and cloud deployment. I love learning new ways to create digital products with Amazon Web Services, and to share with others what I've learned and built.

My Resume

Projects

My Toolkit

AWS Sagemaker
tensorflow
keras
scikit-learn
AWS Lambda
plotly dash
pandas
word2vec
heroku
nltk
AWS API Gateway
xgboost
docker
flask
AWS Textract
spacy
numpy
AWS ElasticBeanstalk
AWS Glue
Tableau
matplotlib
Apache Gremlin
SQL

Projects A few examples of machine learning applications I've built and deployed. I share these with my students as part of my AI/ML course.

Software Screenshot

Deploy an NLP model with SageMaker and Lambda

I deploy a Sagemaker serverless endpoint for a trained NLP model and then connect it to an API Gateway endpoint with a Lambda function.

View the App!

Blog Post

Code on Github
Software Screenshot

XGBoost Digit Classifier

Uses plotly dash, scikit-learn and xgboost to receive handwritten images and predict probability for digit recognition.

View the App!

Code on Github
Software Screenshot

CNN Dogs vs Cats Classifier

A convolutional neural net with tensorflow that receives any image and predicts probability for image classification using the kaggle 'cats vs dogs' training dataset.

View the App!

Code on Github
Software Screenshot

Topic Modeling of Medical Transcripts

I applied Latent Dirichlet Allocation (LDA) to cluster unlabeled medical transcripts into similar topic clusters. Classification accuracy and coherence score evaluate the optimal model.

View the App!

Code on Github
Software Screenshot

Graph Network Analysis

A series of json.gz files representing github connections were downloaded from github archive. Each file was processed and converted to Apache Gremlin format. Gremlin files were then loaded into Amazon Neptune for further analysis.

View the App!

Code on Github
Software Screenshot

Deep Learning for Survival Analysis

This project predicts time til server failure, using a dataset of 125k web servers from 2015. Final analysis employs a neural network model (DeepSurv Neural Cox Analysis) with F1 Score of 96%, ROC-AUC score of 86%.

Code on Github
Software Screenshot

Additional Projects

A blog post showing how to deploy a Sagemaker endpoint for a trained NLP model and then connect it to an API Gateway endpoint with a Lambda function to predict classification of user inputs.

View my other projects

Courses Between 2018 and 2022, I've taught over 500 students to build and deploy machine learning apps on Heroku and AWS. Often I wrote original content; sometimes I repurposed public examples from the web but gave them my personal stamp through modifications and reorganization of materials.

Software Screenshot

Deploying Python Applications on Heroku

This course introduces students to Flask for app-building, Plotly for visualization, and Plotly Dash for building simple ML applications. We use Heroku and Docker to deploy the apps to the web.

Course Materials
Software Screenshot

Deploying Deep Learning Apps on AWS

The course introduces students to tensorflow and keras for trainining and optimizing deep learning models. We use a combination of AWS services (Sagemaker endpoints, lambda functions, API Gateway, S3 static website, ElasticBeanstalk, Chalice) to deploy the trained model to the web for inference. We learn three modalities for deployment: static website (via S3 and API Gateway), web server (via ElasticBeanstalk) and serverless (via Chalice and Lambda).

Course Materials

Certifications A tangible measure of the learning and exploration I've done over the years.
AWS Certfications

  • AWS Data Analytics
  • AWS SysOps Administrator Associate
  • AWS Certified Machine Learning
  • AWS Certified Developer
  • AWS Certified Solutions Architect
  • AWS Cloud Practitioner Certification
  • AWS Certfications

  • ScienceLogic Certified Expert
  • Udacity AWS Machine Learning Engineer
  • Comptia Security+ Cybersecurity
  • AWS Academy Accredited Educator
  • General Assembly Data Science Immersive
  • ScrumMaster Certification
  • IBM Certified Statistics