Zavier Rahmansyah

Projects

Race-Engineer-AI-Assistant, IBM AI Builders Challenge

Currently developing an AI-powered solution that improves team racing for the IBM AI Builders Challenge. The focus area I chose is: Performance Insights. We are expected to Use at least a tool(s) developed by IBM: IBM Granite, Docling, Langflow, Context Forge

Terminal-based AI Agent, Hackerrank Hackathon Project

Developed a terminal-based AI agent for a coding competition. The project was built using Python and Claude API, and it was designed to assist users if they have any questions related to using Claude, Visa, or Hackerrank. The tools, frameworks, and technologies used in this project include: Python and Claude API. Specifically, the program was made using the principles of TF-IDF, not collecting user's data, and providing accurate information to users. The AI model used for the project is Claude sonnet 4.6.

House of Hope Website Redesign, Badger for Good Designathon

Redesigning a hard-to-navigate website for a non-profit organization for a UX Design Competition. Following the competition, I developed a new website for the organization. The tools and technologies used in this project include: Figma, HTML, CSS, and JavaScript.

Hyperstition for Good Writing Competition

Developed a paper for the Hyperstition for Good Writing Competition. The idea behind hyperstition is the research that suggest that writing about AI reasoning with compassion can actually shape how future AI behaves. In other words, we can help influence how AI treats life.

Proximity Match, Claude Builder Club (CBC) Anthropic x College of Computing and Artificial Intelligence

Building an application to help students find their best match for friends, project partners, and other academic collaborations. The project was made using the idea that students may feel uncomfortable to initiate conversations with strangers, but they may be more comfortable to start conversations if they know that they have something in common and if they know they are located within a reasonable distance. The tools, frameworks, and technologies used in this project include: Python, HTML, Javascript, Shell, Vercel, and Claude.

Synchronized Applications Recommender Knowledgebases, Codex Creator Challenge

Currently developing application intended to address the problem when people don't know what they want for the Handshake x OpenAI Codex Creator ChallengeThe tools, frameworks, and technologies used in this project include: OpenAI Codex, OpenAI API, React, Node.js, Express.js Product Hunt API, College Scorecard API, CodeForces API, Figma, Canva, and Model Context Protocol (MCP).

Personal AI Agent: Daily Quest AI Agent

Developed a Personal AI Chatbot Agent using Langchain, React framework for the Single Page Application (SPA), node.js, Express.js, MongoDB for storing data, and FastAPI, Langchain, and OpenAI for the LLM implementation. The tools, frameworks, and technologies used in this project include: OpenAI API, Langchain, React, Node.js, Express.js, MongoDB, and python FastAPI

Personal AI Agent: Morning Briefing AI Agent

Developed a Personal AI Agent to brief my morning with motivation and list of tasks. While I could have just used google calendar and weather app, I wanted to create an AI agent that can integrate information from multiple sources and provide a concise briefing. The project is also beneficial for me to learn how to build an AI agent and how to integrate multiple APIs together. The tools, frameworks, and technologies used in this project include: OpenAI API, Model Context Protocol (MCP) framework, Google Calendar API, and Weather API.

Badger Solar Racing: Race Strategy Team

Working in a subteam of badger solar racing, the goal of the team is to optimize the usage of car's battery to maximize the car's potential. I worked on: Implementing Isolation Forest algorithm to detect outliers in the car's sensor data.

UW-Madison Deliberation Dinners Project

An initiative that brings together UW-Madison undergraduates with diverse ideological perspectives to deliberate on important political issues.

I found this experience to be useful and exciting because not only we get to talk about wide variety of political issues. It also changed my view on how I approach things, meaning to not only standing between "yes" or "no situation", but to actually have a clear standing and idea on where I want to be in this topic. This experience also taught me how I can improve quality of discussions by bringing up one topic over another and how we can further develop a discussion topic when everyone seems to be in a concensus.

Personal Website

This website is built using HTML and CSS, alongside Visual Studio Code, Github, and Netlify. It showcases my background, favorite quotes, projects, work experience, and extracurricular activities.

Project Management: Your Best Choice University Using Multi-Attribute Decision Process

Developed a comprehensive project management paper for selecting the best university based on multiple attributes and decision-making criteria. Worked under Dr. Paul D Giammalvo

Work Experience

International Sibling Society

Data Analyst Intern

Currently working as an undergraduate intern for a non-profit organization focusing on sustainable development goals (SDGs). Possible responsibilities throughout this summer include:

Designed and implemented data collection systems to track user growth, engagement, and retention metrics.
Analyzed platform analytics data using Google Analytics, SQL databases, and similar tools.
Created data visualizations and dashboards using Tableau, Power BI, Google Data Studio, and similar tools.
Generated monthly and quarterly reports tracking KPIs for all four modules.
Designed and administered user surveys to measure satisfaction, usability, and impact.
Conducted qualitative research including user interviews and focus groups.
Analyzed survey and interview data to identify trends, insights, and recommendations.
Supported A/B testing to optimize user experience and conversion rates.
Monitored platform performance metrics such as uptime, page load times, and error rates.
Collaborated with module leads to set data-driven goals and track progress.
Prepared data visualizations and presentations for board meetings and funder reports.