#semanticClimate Intern | Stanford Code in Place Section Leader (Volunteer) | Open Source Contributor | Data & ML System

github

Aleena Harold Peter  • Peerlist

Co-authored an article published on the Open Knowledge Foundation (OKFN) Blog as part of the AI Learning Labs initiative. The article presents the Climate Academy Chatbot, developed in collaboration with semanticClimate, Climate Academy, and OKFN to provide trustworthy, transparent, and citation-backed climate information.It covers the chatbot's Retrieval-Augmented Generation (RAG) architecture, semantic search, grounded responses, source citations, and principles for building responsible AI systems that prioritize transparency, reliability, fairness, and accessibility.My contributions included:<ul><li>Researching responsible AI and climate information systems</li><li>Writing and documenting the chatbot architecture and design</li><li>Explaining semantic search, vector retrieval, and citation-based response generation</li><li>Collaborating with the semanticClimate team to communicate the project's technical approach for a broader audience</li></ul>

A Climate Chatbot for Trusted and Transparent

Built a context-aware AI chatbot for the JobsForHer Foundation platform to assist women in discovering job opportunities, events, and mentorship programs. The chatbot supports multi-turn conversations, detailed job info (salary, skills, links), multilingual FAQs, and user authentication—making the job search experience interactive, inclusive, and efficient. Key technologies included semantic search via SentenceTransformers, retrieval-augmented generation (RAG), and real-time data retrieval using CSV, JSON, and web scraping. Built-in bias detection ensured ethical and inclusive responses. Key Features: Multi-turn conversation support with context tracking Semantic job search using vector similarity (SentenceTransformers) Job info extraction: salary, skills, contract type, job links User login/signup via Supabase Real-time data via CSV, JSON, and web scraping Built-in bias detection for ethical AI behavior

Asha AI Chatbot Hackathon 2025 (🤖 Context-Aw

Built during Walmart Sparkathon 2025 (Theme: Sustainable Retail Innovation), this MVP helps users estimate the environmental impact of common products across different Bengaluru localities. 🔍 Problem: Most consumers don’t realize how daily purchases affect the planet, especially when local factors like energy sources and transport logistics are involved. 💡 Solution Highlights: Simulates emissions across 10+ localities Formula: Product × Adoption Rate × Population Dynamic sustainability tips per product Heatmap visualization using Leaflet.js + OpenStreetMap Arrow-key navigable autocomplete to enhance UX 🛠️ Tech Stack: HTML + CSS (Dark Theme UI) Vanilla JavaScript Leaflet.js OpenStreetMap API

🚀 Carbon Footprint Estimator (Walmart Sparka

Maze of Knowledge is a gamified learning desktop application that transforms traditional quizzes into immersive mini-games. Designed for Java, aptitude, and logical reasoning prep, it blends quiz mechanics with arcade-style gameplay, complete with pixel-art sprites, animated UIs, sound effects, and live coding challenges. ✨Features • 4 Playable Game Modes: ○ Classic Quiz – Timed MCQs with combat animations ○ Matching Tiles – Flip &amp; match question-answer pairs ○ 1v1 Fighter – Quiz answers trigger special attacks (adapted from a JS tutorial to JavaFX) ○ Pixel Retro Runner – Side-scrolling platformer with quiz checkpoints • Live Code Execution: Integrated JDoodle API for Java &amp; Python challenges • Online Leaderboard: Real-time score tracking with Supabase backend • Custom Themes &amp; Audio: Light/dark mode, music/SFX controls, quiz timer settings • Offline Support: Local JSON question cache for uninterrupted gameplay 🛠 Technical Highlights • Built with JavaFX (UI/game logic), FXML/CSS for design, and Gson for JSON parsing • MVC-based modular architecture for easy scalability • API integrations for authentication, leaderboard, and code execution • Lifecycle management to eliminate UI flicker between game modes • Custom sprite animation controller to sync character actions in Fighter mode 📚 Impact / Learnings • Strengthened skills in JavaFX, API integration, and async UI handling • Learned to merge game design principles with educational software • Enhanced ability to debug complex UI/game logic interactions

🧠 Maze of Knowledge – Gamified JavaFX Quiz A

Built a web app that turns simple user prompts into full learning kits PDFs, quizzes, even audio lessons. Added multilingual support, topic classification, and a sleek UI using Streamlit. 💡 Tech Stack: Python, Streamlit, Cohere API, Hugging Face, Google Translate API, ReportLab 📎 Auto-generates resources that are accessible, localized, and scalable.

Pragati Meta Hackathon 2025 – AI-Powered Educ

✨ Features • Interactive Learning Modules ○ Disaster Do’s & Don’ts with region-specific context ○ Educational content tailored for schools and communities • Gamified Quizzes & Leaderboards ○ Quiz-based learning to reinforce disaster awareness ○ Score tracking and competitive leaderboards to boost engagement • Region-Based Emergency System ○ Location-aware emergency alerts ○ Emergency contact directory for quick access during crises • Multilingual Accessibility ○ Supports English, Malayalam, Hindi, and Kannada ○ Designed for inclusivity across diverse user groups • Admin Dashboard ○ Monitor user participation and quiz performance ○ Track preparedness progress at school/community level 🛠 Technical Highlights • Frontend: React.js with Tailwind CSS and Framer Motion for smooth animations and responsive UI • Backend: Node.js + Express.js REST API • Authentication & Security: JWT-based auth with protected routes and middleware • Database: MongoDB (user profiles, progress, quiz data) • Scalable Architecture: Modular backend with controllers, models, and middleware separation • Deployment-Ready: Designed for Railway / Render (backend) and Vercel (frontend) 📚 Impact / Learnings • Gained hands-on experience building a production-style full-stack application • Learned to design education-first systems with real-world social impact • Strengthened skills in API design, authentication, and state management • Understood the challenges of multilingual UX and region-specific data modeling • Improved collaboration and open-source contribution practices Skills: React.js, Node.js, Express.js, MongoDB, REST APIs, Tailwind CSS, UI/UX Design, JWT Auth, Git/GitHub

🌊 Safe and Sound – Disaster Preparedness & R

•	Redesigned interface elements with a linear gradient (#000000 to #3533cd) for a modern look. •	Incorporated Montserrat font for a clean and professional aesthetic. •	Added “Sign Up via Google” and “Welcome Back” messages to enhance user experience. •	Introduced navigation tabs (Quiz Zone, Video Zone, Game Zone, Q&A Zone, Rewards Zone) inspired by YouTube's layout. •	Implemented a timer feature for crossword and word puzzles, and an Achievement System to reward players. •	The prototype of the design was created by Vansh Gupta. The logo of the app is created by Priyanka Suresh.

Smart India Hackathon

🟣 Shortlisted Project | Empowerment Track 🌍 Problem: Rural and semi-urban women face limited job access, weak digital skills, and lack business training. 👩💻 Solution: Designed “StreeBal,” a mentorship-first digital training initiative to upskill women entrepreneurs. Built clickable HTML prototypes showing learning modules, support programs, and community hubs. 📑 Deliverables: Proposal + Interactive landing pages + Implementation strategy. 🔗 Aiming to turn grassroots resilience into entrepreneurial growth.

💪 StreeBal – Udyam Mahila Ideathon 2025

Uplift is a mobile and web application built with Flutter to help students from lower-income backgrounds discover, track, and apply for scholarships. It supports multi-language accessibility (Hindi, Kannada, English), enabling inclusive access to financial aid opportunities. Core features include scholarship filtering, bookmarking, deadline reminders, and a basic chatbot to guide users. The platform promotes accessibility, equity, and is designed for scale, with a roadmap that includes AI-powered recommendations and expanded language support. Key Features: Scholarship filtering with cached CSV data Local notifications with live countdowns Application tracking & bookmarking Community forum for peer support Multi-language UI with Google ML Kit Basic rule-based chatbot for FAQs

Uplift — Scholarship Discovery & Tracking App

•	Developed a menu-driven program with inheritance for department-specific CGPA and SGPA calculations. •	Included robust error handling using try-catch blocks for user input validation. •	Classes structured hierarchically with a base class (Department) and derived classes (CSE, CSE_AIML, ISE, Aeronautical, ECE).