Analytics website for energy dataAWS: ElasticBeanstalk, EC2, S3, Cloudwatch, Codecommit, CodeDeploy, CodePipeline, Route 53, and LambdaPython: Dash, Pandas

-Created automated QA/QC and normalizing data-pipeline for qPCR RNA tissue data, custom chemistry-Created Generative model based on Neural-Nets for purpose of demonstrating class differences between Immune-Checkpoint-Inhibitor drugs of Responder, Nonresponder, and Undiseased of simulated spatially collected tissue samples of cancer patients-Created dashboard for purpose of inputting custom data to Generative model, viewing training progress, and sampling of generated images at epoch intervals, created differential gene RNA expression plots based off unique data clusters discovered from breast cancer data-Data normalization, distribution validation, dimensionality reduction, hierarchical clustering, K-means clustering, and novel gene co-expression analysis with consideration of ROIs (tissue regions of interest) to identify genetic patterns between the data. Co-expression analysis took into account unique value percentages, correlation confidence interval thresholds, and positive/inverse correlation thresholdsMethod specifics: Generative model built with WGAN architecture, discriminator built with xception CNN architecture. Stratified data validation with MW-U testing. Distribution normality AB testing with Shapiro-Wilks. Graphical based analysis with gene co-expression networks with networkx and and visualized with gephi

-Created Recommender system that takes into account 5 different user selected parameters for movies, restaurants, and alcohol choices-Integrated facial recognition APIs from AWS and Azure to demo validity and performance for use in dangerous person recognition-Business research, outreach, and technical auditing to connect machine learning startups with Ryver, leading to Defendry. Additional QA/QC for PoC deployment of the platform-PoC development work with AWS IoT to integrate to a secure front-end for management and deployment with Raspberry Pi and Arduino devices-Interim devops planning and management of LeadMD on AWS with k8s

Working with data in the form of drawings and specifications to determine a project price and time line.

Using data in the form of drawings and specifications to determine project price and time line.

Takeoffs, submittals, Subcontractor quotes, Supplier quotes, RFIs, Jobsite inspections