Tirus Kimani Wagacha,Data Science & Generative AI Portfolio

About Me

I am a data scientist and software engineer focused on building intelligent, data-driven systems that deliver measurable business impact. My work sits at the intersection of scalable backend engineering, machine learning, and generative AI. I design systems that automate insight generation, reduce operational friction, and empower data-driven decision-making.

I hold a Master of Science in Data Analytics-Data Science from Kansas State University (GPA: 4.0) and bring hands-on experience across fintech, logistics, and insurance domains. My core strengths include building LLM-powered agents, NLP pipelines, and self-serve analytics platforms, as well as deploying production-ready solutions that scale reliably.

I enjoy translating complex data into clear, actionable intelligence and building systems that help organizations move faster, smarter, and with confidence

Download Resume See Projects

Chat with Me

Ask about my work, projects, or how I build intelligent systems.

Skills & Technologies

Combining analytics, engineering, and AI to build intelligent systems.

Python

Data manipulation, ML pipelines, automation, and agent logic.

SQL & NoSQL

Complex queries, data modeling, and scalable storage (MongoDB/MySQL).

Visualization & Storytelling

Power BI, Tableau, narrative-driven insights for decision-makers.

Machine Learning & LLMs

Predictive models, transfer learning, LangChain, GPT-4 integrations.

Selected Projects

Real-world solutions combining data, AI, and systems engineering.

Tableau

Analyzing Subsidies & Emergency Support for Livestock

Explored government agricultural subsidy distribution across Kansas, Nebraska, and Oklahoma to surface inefficiencies and strategic positioning for livestock farmers using Tableau.

Read Article

Data Storytelling

The Pricing Dilemma on Amazon

Investigated discount strategy effects on customer trust and engagement, balancing perceived value and ratings using Amazon sales data.

Read Article

Power BI

Analyzing Survey Data

Transformed and visualized responses from 600–700 data professionals to surface trends around compensation, job satisfaction, and tooling preferences in the data industry.

Read Article

Comparing Writing Style Across Reuters Regions

NLP + ML

Comparing Writing Styles: Regional Variation Between U.S. and European News Articles

Project to extract, clean and analyze writing style differences between Reuters articles by region (US vs Europe). The analysis pipeline extracts articles, computes linguistic features (POS,Sentiment, Embeddings), runs topic modeling and classification, and outputs CSVs and per-article text files.

View Project

NLP + ML + GenAI

Semantic Question Extraction & Template Builder for Survey Optimization

This project automates consolidation of disparate historical survey questions into a concise, validated generic survey template. It combines semantic embedding similarity, clustering and topic modeling, and human-in-the-loop review with LLM-assisted paraphrasing to produce high-quality, reusable questions and an Interactive Streamlit Dashboard.

View Project

NLP + ML

Sentiment-Enhanced Product Recommender System

A sentiment-enhanced recommendation system was built using fine-tuned BERT models on SST-5 for sentiment analysis and FAISS for similarity searches. It combines sentiment scores from customer reviews with product embeddings to deliver tailored recommendations.

View Project

Text Analytics

Customer Feedback Analysis for Hospitality

Analyzed hotel reviews to help AfriDusky Tours & Travel Agency enter the Kenyan market with customer-centric insights on satisfaction, strengths, and improvement areas.

View Project