{"id":29,"date":"2021-11-29T03:56:29","date_gmt":"2021-11-29T03:56:29","guid":{"rendered":"https:\/\/shayhmaor.dev\/?page_id=29"},"modified":"2026-05-01T15:01:13","modified_gmt":"2026-05-01T15:01:13","slug":"about-3","status":"publish","type":"page","link":"https:\/\/shayhmaor.dev\/","title":{"rendered":"Shay Maor – Profile"},"content":{"rendered":"\n

Shay Maor is a Senior Level Developer with over 12 years of Information Technology experience. He has 10+ years of software development experience including over 10 years Hands on Industry Experience and 6 months Academia Development experience as a Member of a Center for Computational and Integrative Biology Research Group. He has experience in Microsoft C#\/.NET (8+ years), Python (2+ years), and Java (1 year). He graduated from Rutgers University in May 2020 with a Bachelor of Science in Computer Science and a Minor in Psychology with a focus on Software Development electives and 2 Computational Biology Independent Studies. He graduated with a Master of Science in Software Development from Boston University (ranked #8 in Online Information Technology Programs by US News) taught by MET Computer Science Faculty with a focus on BU MET Computer Science Electives in 2025.<\/p>\n\n\n\n

Content Management Systems<\/strong> Experience<\/strong><\/h2>\n\n\n\n

Highlighted work includes Search Engines, Commerce Functionality including Wishlist and SEO Metadata, Reporting Automation, Conference Reporting, Responsive Data Visualizations, Custom Content Migrations of over 20,000 articles between different CMS, Custom Webhook APIs, Responsive Upgrades, and much more.<\/p>\n\n\n\n

AI Experience<\/strong><\/h2>\n\n\n\n

Built a diverse and highly technical portfolio spanning enterprise generative AI, computer vision pipelines, and on-device machine learning. Engineered a robust LLM and NLP engine for the legal sector using Full Stack .NET technologies, SignalR, and Telerik, successfully integrating OpenAI and Mistral to automate the complex analysis and processing of legal contracts. Designed sophisticated Python data pipelines capable of digitizing static mathematical equations into functional LaTeX code, as well as extracting chemical intelligence by translating visual molecular structures into machine-readable SMILES formats and spatial MOL representations via an experimental RDKit module. Additionally, developed Bookoo AI Brew Assist: a localized, GPU-accelerated iOS application leveraging small-parameter LLMs and real-time telemetry to diagnose and optimize espresso extraction.<\/p>\n\n\n\n

Portfolio<\/strong><\/h2>\n\n\n\n
\"\"
Metabolic Optimization and Simulation Tool<\/strong>
Java based systems biology tool designed to research organisms and their respective metabolic networks, with the intention of both simulation and discovering ideal gene knock-out combinations for increasing production of target metabolites. Originally as part of a Computer Science independent study. Added several components to pre-existing software package.<\/figcaption><\/figure>\n\n\n\n
\"\"
Machine Learning: Psychology and Movie Preferences<\/strong>
Senior design project that implemented data mining techniques and a psychological assessment test (MBTI) to discover correlations and predictive models on movie preferences. Primarily, using numerical quantification of the 16 cognitive functions underlying the MBTI to find correlation with sub-genres using a J48 decision tree-inducing algorithm under Weka Machine Learning software. The J48 algorithm is Weka’s implementation of the C4.5 algorithm by Ross Quinlan.<\/figcaption><\/figure>\n\n\n\n
\"\"
Molecular Dynamics Independent Study<\/strong>
Under a Physics Independent Study, worked on simulating and analyzing coarse grain models of lipid bilayer membranes using Molecular Dynamics software packages (GROMACS, NAMD, VMD) and TCL. The main focus was the recreation of the simulation in the study “Phase Separation in a Lipid\/Cholesterol System: Comparison of Coarse-Grained and United-Atom Simulations.”<\/figcaption><\/figure>\n\n\n\n
\"\"
Open Source Genome<\/strong>
OS Genome is an open source web application that allows users to gather the information they need to make sense of their own genome without needing to rely on outside services with unknown privacy policies. OS Genome’s goal is to crawl various sources and give meaning to an individual’s genome. It creates a Responsive Grid of the user’s specific genome. This allows for everything from filtering to excel exporting.<\/figcaption><\/figure>\n\n\n\n
\"\"
OSGenome v2<\/strong>
OSGenome v2 is a high-performance, privacy-first open-source web application that empowers users to locally cross-reference their raw 23andMe DNA data with the SNPedia database. Re-architected to handle massive datasets of over 25,000 variants instantly, it leverages client-side deferred rendering to eliminate browser lag and features a rebuilt, clinically accurate parsing engine that guarantees zero data loss when handling complex heterozygous indels and single-allele traits. Beyond robust data processing, v2 introduces categorization, generated by an offline, low-parameter local AI engine that “intelligence-maps” the genomic library variants with discernible traits, injecting interactive, searchable biological categories directly into the lightning-fast UI so users can securely explore their genetic footprint with unprecedented speed and insight.<\/figcaption><\/figure>\n\n\n\n
\"\"
SENT-PHARMA (Data Visualization Layer)<\/strong>
Concurrent Pharmaceutical Sentiment Analysis + Stock Valuation Tracker in Java Spring for my Boston University class, Advanced Programming Techniques. With several key MVC Endpoints for news aggregation, sentiment analysis using Microsoft Azure Cognitive Services, and Stock Valuation time series retrieval. Pictured above is the clustering D3.js data visualization using an MVC Endpoint for data population.<\/figcaption><\/figure>\n\n\n\n

<\/p>\n\n\n\n

\"\"
SENT-PHARMA (Prediction Algorithm Layer)<\/strong>
Python script leveraging SQL Alchemy to aggregate and analyze data from the SQL Server I set up on AWS Linux Cloud and a Neural Network using Linear Regression and Sigmoid Function to predict stock market behaviors. Pictured above is the Prediction Algorithm predicting loses for Takeda and Pfizer, and gains for Merck. Takeda had losses after hours, with the prediction accounting for news aggregated after market hours<\/figcaption><\/figure>\n\n\n\n
\n