Trupti Agrawal TruptiAgrawal

✨ Trupti Agrawal ✨

she/her · IT student @ BVRIT Hyderabad · Hyderabad, India 🌸

🔭 Currently exploring DSA, systems programming, and applied ML
🌱 Documenting everything I learn in my Learning_log — syntax, algorithms, and little "aha!" moments across languages
⚙️ Built Futoshiki, a constraint-based logic puzzle solver implemented across 4 languages (Go, Elixir, Rust, Python) to compare how different paradigms shape solver design
🍵 Brewed taskBrew, a full-stack AI task manager where you type tasks in plain English and an LLM parses them into structured tasks — Spring Boot + JWT on the backend, React on the front
🌿 Building SAGE, an ML pipeline that predicts LLM token usage & response quality from prompt features, comparing cost/quality tradeoffs across models
🚌 Modeled real-world college transportation as a Capacitated Vehicle Routing Problem in Bus Route Optimization
🖥️ Co-building Nirmaan, a gamified platform teaching Git & Linux through interactive simulations — 🥈 2nd Place, R&D Showcase (SDG 4: Quality Education)
♟️ Built ChessGame — a fully playable chess game using functional programming principles with a Tkinter GUI
📝 Contributed docs fixes to mem0ai, Slicer and freeCodeCamp
🦈 Proud holder of the Pull Shark achievement
🏆 750+ problems solved across LeetCode, CodeChef, Codeforces & HackerRank · 50+ contests
💌 Reach me at truptiagrawal777@gmail.com

🧩 Futoshiki A constraint-propagation + backtracking solver for the Futoshiki puzzle, built in 4 languages with interactive step-by-step visualizations of every solver decision. `Go` `Elixir` `Rust` `Python`	🖥️ Nirmaan A gamified developer-learning platform with a Git state-machine simulator (20+ commands, DAG-backed) and a browser-based Linux terminal emulator (40+ commands). `React` `Git` `Linux`
🍵 taskBrew A cozy full-stack task manager with JWT auth, projects & tasks, and an AI assist that turns plain-English input ("finish the report by friday, high priority") into structured tasks via an LLM. `Java` `Spring Boot` `React` `PostgreSQL`	🌿 SAGE Smart Advisor for Generative modEls — an end-to-end ML pipeline predicting LLM token usage & response quality from prompt features, benchmarking RandomForest, CatBoost & LightGBM to map the cost/quality tradeoff across models. `Python` `scikit-learn` `CatBoost` `LightGBM`
🚌 Bus Route Optimization Modeled transporting 690+ students across 15+ buses as a Capacitated Vehicle Routing Problem, cutting unnecessary travel distance with an optimized routing pipeline. `Python` `uv` `CVRP`	♟️ ChessGame A fully playable chess game built with functional programming principles and a Tkinter GUI — clean, modular, and fun to play. `Python`

✨ thanks for stopping by ✨