Hire the world's best freelance Amazon EC2 experts

During my years as a researcher in both academia and the corporate sector (R&D) I have tackled problems in many domains.  I have recently transitioned back to academia as a researcher at the University of Michigan, but my experience over the past couple years in a biotech startup led to me successfully automating and scaling a compute cluster, integrating automated workflow pipelines, building NGS clinical carrier screening software that passed CLIA certification, along with providing linux administrative help to fellow bioinformaticians.  Basically, I'm used to wearing multiple hats and at the end of the day I enjoy solving problems.  If you have an interesting or pressing problem that needs someone dedicated that will see it through to solution I hope to be your man.

Below are a few examples of finished projects:

Data Munging and ETL Processing - The foundational work of any data science project or platform

• The largest project that I completed in this area was at IHBI where I implemented an automated Greenplum (now called Pivotal HD) ETL tool.  Written in Python, the tool automates the transformation of differently-formatted input data while managing balanced dispersion across all available database nodes via templated SQL queries.  The Pivotal interface between SAS and Greenplum had major trouble with balanced distribution, necessitating this proprietary solution.

Predictive Analytics - A few of my best individual efforts in predictive analytics were submissions to the Kaggle data science competition platform.

• Built a NN model for the Liberty Mutual Group: Property Inspection Prediction challenge that got me a bronze medal and placed me in the top 10%
• In the Grasp-and-Lift EEG Detection competition I built and trained an extreme gradient boosting model that got me a silver medal and placed in the top 12%
• My submission to the original March Machine Learning Mania competition used a combinatorial approach with an ensemble of multiple nonlinear regression models that placed in the top 15% and beat all benchmarks, including the "Wisdom of the Crowd"-esque metric

Deep Networks - A key interest area I have been studying since 2013 and have continuously applied them in the bioinformatics domain to a myriad of problems.  This is still a rapidly evolving area and I enjoy keeping up on every new insight.

• Built software that automated neural network hyperparameter optimization (meta-learning) and subsequent training that led to a Maxout NN model used in a tiered approach in a review of protein disorder prediction for The International Journal of Molecular Sciences