eAxis Analytics is available for a broad range of independent consulting and contracting
services covering Big Data ETL pipelines, data processing, analytics, and modeling.
eAxis Analytics was founded by Michael Suesserman to develop novel AI models for
automated stock analysis. Michael possesses a PhD in Electrical Engineering with
a major in Digital Signal Processing that covers a broad range of advanced theoretical
and applied mathematics. A passion for working with data underlies over a decade of
experience processing and analyzing many types of structured and unstructured data
with an emphasis on financial, ecommerce and biomedical fields. Complete details are
available on LinkedIn:
Projects:
Recent projects focus on three general areas of Deep Learning:
-
Image Processing and Computer Vision: currently training a custom
neural network to perform technical stock analysis by reading and interpreting images
of stock charts as a means of predicting stock prices. Project utilizes OpenCV, CNN,
Object Detection, Mask R-CNN, and Keypoint Detection.
-
Time Series Analysis: trained a Recurrent Neural Network (RNN) with
LSTM to extract information from stock data for predicting prices. Also, built a Big
Data ETL pipeline on AWS for ingesting and analyzing stocks in real-time. Project
utilizes RNN with LSTM, Restricted Boltzmann Machines, Deep Belief Networks (DBNs)
and AWS (Kinesis, Glue, Redshift & S3).
-
Natural Language Processing (NLP): analyzing public companies using
NLP technologies for use in predicting stock prices. Project utilizes tokenization,
NER, topic modeling, semantic analysis and sentiment analysis.
Contracting Services Include:
- Data Engineering --> ETL, Data Pipelines and Data Lakes
-
AWS (Kinesis, EMR, Glue, Redshift & S3)
-
NoSQL (DynamoDb, mongoDB & Cassandra)
-
Apache Spark (Scala & Python)
-
Hadoop Ecosystem (HDFS, YARN, MapReduce, HBASE, …)
-
SQL (MySQL, Postgres and SQL Server)
- Data Science:
-
Data Processing
- Predictive Analytics and Machine Learning:
- Supervised --> regression and classification
- Unsupervised --> clustering
- Deep Learning:
-
Supervised --> Artificial Neural Network (ANNs), Convolutional Neural Networks (CNNs),
Faster R-CNN, Mask R-CNN, Recurrent Neural Networks (RNNs) with LSTM
-
Unsupervised --> Self-Organizing Maps, Boltzmann Machines, AutoEncoders, Deep Belief
Networks (DBNs)
- Artificial Intelligence:
-
Reinforcement Learning --> Deep Convolutional Q-Learning, A3C, LSTM-A3C
-
Natural Language Process (NLP) --> tokenization (POS, lemmatization, …), NER, topic
modeling, Word2Vec, semantic analysis and sentiment analysis
- Commonly Used Languages, Libraries and Platforms:
-
Primary Languages/Platforms (used daily): Python and Anaconda3 (including numpy, matplotlib, pandas, sklearn, tensorflow, keras & pytorch) on LAMP stack.
-
Other Languages/Platforms: ASP.NET/MVC, C#, C++, PHP, Node.js/jQuery and client technologies (JavaScript/JSON, HTML, CSS, XML)
-
Cloud Services: AWS including SageMaker and AI Services (advanced expertise) and Azure (basic experience)
-
Development Stacks: Microsoft and LAMP
-
Databases: SQL (MySQL, Postgres & MS SQL Server) and NoSQL (DynamoDb, mongoDB & Cassandra)
-
Microservices: Amazon AWS Lambda
-
Containers: Dockers (with Flask for web apps)
-
Visualization and Dashboards: matplotlib, Jupyter, Tableau and R-Studio
-
Computer Vision: OpenCV
-
NLP: Spacy, Gensim and NLTK
eAxis Analytics Web Sites
eAxis Analytics currently operates three separate sites all running in the cloud:
-
Main Site
(eAxis.com):
this corporate site that summarizes our Data Science Consulting Services.
-
Serverless Dev Site
(dev.eAxis.com):
a dev site that is used for developing and testing serverless microservices on AWS (Amazon
Web Services). This entire site is created using static HTML that is globally available
through the CloudFront edge network. Dynamic content is created using AWS Lambda and
Lambda@Edge connected to the API Gateway.
-
Machine Learning Site
(ml.eAxis.com):
a server-based site that is used for developing and testing machine learning, deep
learning and AI models that are too complex to run as serverless apps. This site is
scalable and highly available running on EC2 Linux servers with load balancing and auto scaling.
I’ll be adding a blog soon with lots of content including articles covering how I developed
and deployed the above sites as well as detailing the public demos included on these sites.