My Projects

Android + iOS

WhimAway: Capture → Tasks

A shipped task-capture product with native Android and iOS apps. Capture tasks through text, photos, or voice, then review, organise, and keep them in sync through one shared backend.

KotlinJetpack ComposeSwiftSwiftUIRoom (Offline Cache)WorkManagerFirebaseFirestoreCrashlyticsFastAPIOpenAI (Multimodal)RetrofitCI/CD (GitHub Actions)Railway
web

Tech Developer · MAYA NPO

Led the end-to-end rebuild of MAYA’s official website, transforming it from a restrictive no-code setup into a high-performance, fully custom Next.js platform. Designed a responsive, animation-rich user experience with a clean, modular architecture that allows non-technical volunteers to manage and update content independently. Reduced annual technology costs by ~75% through infrastructure optimization and smart tooling decisions, directly increasing funds available for child welfare initiatives.

Next.jsTypeScriptTailwindCSSFramer MotionVercelUI/UX DesignPerformance Optimization
research

Game-Theoretic Fairness in Federated MEC Systems

Co-authored a comprehensive academic survey on CAMARA Operator APIs for Federated Multi-access Edge Computing (MEC) services within the emerging domain of Game Theory and Economic Paradigms (GTEP). Explored the strategic and incentive-driven dynamics of resource federation among MEC operators, emphasizing fairness and truthful participation in distributed markets. Conducted an in-depth analysis of API categorization, real-world telecom implementations, and interoperability challenges. Personally tested public sandbox APIs from operators like Orange and Vodafone, and developed demo applications to validate allocation and settlement mechanisms. Paper currently under peer review for an A* conference in an international venue.

ResearchGame TheoryMulti-Agent SystemsFederated MECRESTful APIsCAMARAOpen SourceSandbox Testing
android

On-Device RL: Tic Tac Toe

Engineered a novel variant of Tic-Tac-Toe that introduces strategic uncertainty — the player’s symbol (X or O) changes randomly every turn, removing deterministic advantage and forcing adaptive decision-making. Designed and developed a complete Android app featuring an embedded Reinforcement Learning (RL) agent that learns and plays in real time, entirely on-device. The agent was implemented using three distinct paradigms — Rule-Based Logic, Q-Learning, and Monte Carlo Tree Search (MCTS). The system adapts policies in real time, balancing exploration and exploitation under probabilistic conditions. Demonstrated a consistent >70% win rate against human players, highlighting the potential of lightweight, on-device reinforcement learning without any cloud inference.

KotlinXMLReinforcement LearningQ-LearningMonte Carlo Tree SearchRule-Based AIOn-Device ML
web

Interactive Network Simulator for Classrooms

Developed a web-based simulator to visualize classic networking topologies like Butterfly, Multistage Interconnection Networks (MIN), and Delta Networks. Aimed at making abstract telecom and computer network concepts intuitive through clean, interactive graph visualizations. Officially endorsed and used by IIT Kharagpur professors to demonstrate network routing principles to B.Tech and M.Tech students during classroom sessions.

HTML5CSS3JavaScriptD3.jsCanvas APIGitHub Pages
web

Gamified Student Progress Tracker

Designed a playful, goal-oriented progress tracker to help students boost productivity through gamification. Users can set tasks with difficulty levels and assign XP points. Completion triggers achievements, confetti celebrations, and level-ups; bringing a fun, rewarding feedback loop to daily progress. All task data persists locally for seamless recurring use, making it perfect for long-term habits and academic sprints.

HTML5CSS3JavaScriptLocalStorageCanvas APIGSAP
android

Apps by saipranith.dev

My personal Android development label on the Google Play Store, where I publish WhimAway and Pic-Pac-Poe.

Android StudioJavaKotlinXMLPlay Store ConsoleFirebaseRESTful APIsOpen Source APIs