Alexander Benasutti Email & Phone Number

Middletown, Nj, Us

Middletown, New Jersey, United States

El Segundo, California, United States

Developed an application and database to ingest and organize 10 million+ IP addresses using Python and Azure Cosmos DB, used to identify and reclaim unused IPs in a $100 million cost-savings effort.Created detailed summary statistics visualizations in Cortex XSOAR and Apache Superset using display automations connected to an Azure CosmosDB PostgreSQL database to display various metadata concerning those millions of IP addresses.Designed, developed, and deployed preprocessor and OCR engine solutions to analyze 600,000+ contract documents using Python, highly improving overall accuracy of input document readability.

Ewing, Nj

Assisted in the development of Machine Learning projects relating to image classification and segmentation using DCNNs and SSDs.Worked on building an autonomous rover that will classify, pinpoint, and exterminate weeds in real time using neural networks. Responsible for integrating image segmentation onto the rover.Visualized and conducted analysis of feature maps produced by a trained discriminator network to determine what allowed the network to trivially discriminate between natural and synthetic images of grass.Part of a tight-knit team that divides itself into hardware and software sub-teams, allowing for members to contribute to a spectrum of tasks ranging from model creation to rover motor control.

Ewing, Nj

Researched and presented a novel approach to automatic classification of different colorspace conversion matrices directly from luma/chroma pixel data using DCNNs.Worked with trained colorspace classification models that used the Apache MXNet framework and the ImageNet dataset converted to both Rec. 601 and Rec. 709. Analyzed models for clipping that would occur if the wrong matrix was used, tested validation accuracy when high clipping was largely eliminated from frames of standard video processing datasets.Created DCNN analysis utilities in Python for plotting training progress and dumping feature maps per image at a selected convolutional layer. Created montages of the 16- ‘best’ images (at exciting positive activations) for each feature map. Former has been made available to the public on the MXNet forum and latter is in the process of optimizing for publication.Poster presented in September 2018 at MUSE 2018 Poster Session and at AIPR 2018 conference in Washington DC.

El Segundo, California, United States

Parameterized and generalized Bayesian Networks to automatically detect the most probable root cause of VoLTE network outages by call data record analysis.Developed a Flask API to automatically create and query Bayesian Network models, designed to be callable automatically by Camunda workflow or via chatbot/UI. The API was deployed using Docker and Kubernetes.Utilized Terraform to develop a repeatable pipelined infrastructure using a jumpstart kit and reconfigured according to developer specifications, such as implementing a PostgreSQL database.Researched and presented methodologies for pointing to an AT&T "golden image" in Terraform, as well as a comparison of utilizing Virtual Machines to Virtual Machine Scale Sets, and whether one would be more applicable to the team regarding their application.Developed a web application that assists healthcare professionals, prospective PPE donators, and incoming patients that helps consolidate hospital information and facilitate donations to those most in need.

Greater New York City Area

Designed and developed a detailed visualization tool for over 1600 Lightcloud Embedded Linux Gateway sites around the world using the Google Maps Platform. The tool allows employees to locate company information per site, cluster and filter sites based on Gateway health, autocomplete search by category, and use a saving feature to keep track of company and device information.Analyzed prototype versions of Lightcloud products and wrote comprehensive failure cases to be reported back to the company’s offshore team.Assisted in manufacturing engineering of over 200 Digi XBee Industrial Gateways.To complete these projects, I used a combination of Python, Javascript, CSS, HTML, PHP, and SQL to learn how to create a web app, query a PosgreSQL database, and create test cases for firmware version control using TestRail.Participated in software development team kanban stand-up meetings and full team meetings discussing current projects and prospects for further progress.Presented demonstrations for the Lightcloud sites Google Map and led a code review to “pass the torch” onto other developers.

Vincentown, Nj

Assembled and soldered circuit boards using schematics, various electronic devices and data communications cable assemblies for integration into self-service retail touch screen ordering systems.

Bachelor Of Science - Bs, Computer Engineering

Activities and Societies: Intelligent Media Processing Laboratory (IMPL) Member, Institute of Electrical and Electronics Engineers (IEEE).

Education record