NEXUS AI
Our Projects
Explore a selection of real-world projects that showcase our work at the intersection of AI/ML, data science, and leadership strategy. From predictive models and neural networks to RAG-based assistants and value-mapping tools, these examples demonstrate how we build ethical, scalable solutions that solve complex problems and drive measurable impact.
Internal Examples:
01
Worldview Explorer
Built an interactive app that helps users explore their values and beliefs through psychometric scoring, clustering, and historical figure alignment powered by generative AI.
Key Achievements:
-
Generated synthetic worldview profiles using GenAI for over 600 public figures
-
Developed scoring algorithms and clustering system to place users into belief-based groups
-
Built PCA/UMAP visualizations to map ideological proximity and divergence
-
Integrated narrative generation and personalized reports for educational and civic use
02
ORBIT - RAG
Developed the ORBIT-RAG (Organizational Readiness and Behavioral Insights Toolkit – Retrieval-Augmented Generation) application using Streamlit to empower users to query curated civic media documents through natural language. The app blends semantic vector search with generative AI to deliver accurate, context-aware responses tailored to change management and leadership challenges.
Key Achievements:
-
Architected a scalable document ingestion and embedding pipeline using LangChain and ChromaDB for fast, relevant retrieval
-
Built a clean, intuitive Streamlit interface supporting interactive, real-time exploration of organizational and civic content
-
Integrated OpenAI’s GPT models to generate coherent, context-grounded answers from retrieved documents
-
Enabled retrieval-augmented generation from a vetted civic media corpus, supporting informed decision-making and AI-augmented strategy development
03
Easy Visa
Built predictive and clustering models to identify key factors influencing U.S. work visa approvals, enabling faster case review and applicant profiling for immigration agencies.
Key Achievements:
-
Achieved high recall performance on test data using optimized decision tree classifiers
-
Engineered visualizations and stakeholder-facing dashboards to communicate insights
-
Conducted advanced EDA including Cramér’s V, t-tests, and Cohen’s d for feature relevance
-
Applied KMeans clustering to segment applicant profiles based on visa likelihood and characteristics
04
AI Academic Reading Tool
Created innovative AI-enhanced reading tools for 2,000+ undergraduate students using fine-tuned LLMs on Azure cloud infrastructure.
Key Achievements:
-
Built real-time analysis dashboard for professors
-
Implemented ethical AI framework for bias detection
-
Designed adaptive content improvements based on usage analytics
-
Integrated privacy-preserving data collection methods
