{"slug": "what-s-for-dinner-aaj-kya-banega-a-20b-model-trained-to-solve-my-mom-s-kitchen", "title": "What's for Dinner/Aaj Kya Banega? A 20B model trained to solve my Mom's kitchen crisis without the chatty fluff", "summary": "A developer built \"What's for Dinner?\", an open-source meal-planning assistant that suggests up to three Indian home-cooked dishes from ingredients on hand, by fine-tuning the 20B open-weight gpt-oss-20b model with the Tinker Cookbook into an adapter called mom-chef-v1. The adapter, hosted on Tinker's cloud GPUs, was trained on a synthetic dataset to give atomic, decisive responses rather than chatty validation, and the whole fine-tuning run cost $0.16. The React and FastAPI app uses the Backboard API to remember recently eaten meals and is deployed as a single web service on Render.", "body_md": "*This is a submission for the [Hacktoberfest Weekend Challenge: Build for a Friend](https://dev.to/challenges/hacktoberfest-weekend-2026-10-01)*\n\nThe most dangerous question in any Indian household:\n\n```\n\"Aaj khane mein kya banau?\"\n(What's for dinner?)\n```\n\nObviously, most days I'll say, \"Make whatever you want.\" And there's a good chance either Mom gets frustrated with that answer, or we're having the same dish for the fourth time that week.\n\nSo I built **What's for Dinner?** for my Mom — and honestly, for the survival of the whole family (especially me).\n\nIt's an AI-powered meal-planning assistant that takes **what's actually sitting in the fridge**, using the names we naturally use at home — *palak*, *paneer*, *bache hue chawal*, etc. — and suggests at most **3 Indian home-cooked dishes** that can actually be made.\n\nIt also remembers what the family has eaten recently, so nobody has to suffer through *Aloo Gobi* three days in a row.\n\nBut the most important feature isn't the recipes.\n\n**It's knowing when to stop talking.**\n\n**Stop asking. Start cooking.**\n\n*The demo is running on Render's free tier, so the first request may take a little longer while the service wakes up.*\n\nThe entire project is open source:\n\nA full-stack application that suggests meals based on your available pantry items. It leverages the **Tinker API** to sample LLM-generated suggestions and uses the **Backboard API** to maintain a history of your recently selected meals, ensuring you don't get the same suggestion twice!\n\n`main.py``/api/suggest` for meal suggestions and `/api/select` to save your choice.`frontend/`` train/``train.jsonl`).` render.yaml` & `build.sh`\nThe React frontend and FastAPI backend are deployed together as a single Web Service on Render, making the deployment simple and helping me make the most of the Hacktoberfest credits.\n\nI wanted to build something **fast, lightweight, culturally aware, and actually useful in a real Indian kitchen.**\n\nI fine-tuned the **20B open-weight `gpt-oss-20b` model** using the Tinker Cookbook.\n\nI trained an adapter called **`mom-chef-v1`** on a synthetic dataset containing:\n\nThe goal wasn't just to teach the model about Indian food.\n\nI wanted to teach it **how to respond**.\n\nThe adapter is hosted on Tinker's cloud GPUs, so my application doesn't need to carry around ~43 GB of model weights.\n\nAnd the best part:\n\n**The entire fine-tuning run cost me just $0.16.**\n\nI integrated **Backboard** as the family's memory layer.\n\nIt keeps track of recent meals and provides that context to the model before it makes a suggestion.\n\nThis means the AI isn't only looking at what's in the fridge.\n\nIt also knows what we've already eaten.\n\nSo if someone says:\n\n```\n\"Aloo ki sabzi toh kal hi bani thi.\"\n```\n\nThe system can treat that as a real constraint instead of suggesting the same thing again.\n\nThe React application is compiled into static assets and served directly through the FastAPI backend.\n\nEverything is deployed as a **single Web Service on Render**.\n\nThe deployment configuration is included in the repository in `render.yaml`.\n\nThis project started because I tried using generic AI models for my Mom. They understood Hinglish and local ingredients, but they completely failed to understand *how* she wanted the answer. \n\nA typical interaction looked like this:\n\n**Mom:** \"Aaj kya banau? Aloo hai, paneer hai aur palak bhi hai.\"\n\n**AI:** \"Aloo ki sabzi sounds wonderful!\"\n\n**Mom:** \"Aloo ki sabzi toh kal hi bani thi, koi nahi khayega.\"\n\n**AI:** \"Yes, you're absolutely right! Since you've already had aloo yesterday, let's explore some other delicious alternatives...\"\n\nMom doesn't want a conversation. She doesn't want validation. She wants the AI to eliminate the bad option and give her a practical answer immediately.\n\nSo, I designed *What's for Dinner?* around **atomic, decisive responses**. Instead of a chatty assistant, it just gives the options:\n\n```\n Palak Paneer  \n Paneer Paratha \n Kadhai Paneer\n```\n\nNo explanations. Just the answer.\n\nThis is where **open innovation** changed the game. Instead of fighting a closed model's inherently chatty nature with massive, brittle system prompts, I fine-tuned an open-weight model on household constraints and the exact short, decisive answers my Mom expects.\n\nThe goal wasn't just to build an AI that knows Indian food.\n\n**The goal was to build an AI that knows when to stop talking.**\n\nAnd thanks to open weights, I was able to build it for exactly **$0.16**.\n\nI'm submitting What's for Dinner? for:\n\n**Best Use of Tinker** — Fine-tuned gpt-oss-20b into mom-chef-v1 for concise, culturally-aware Indian meal suggestions. Fine-tuning cost: $0.16.\n\n**Best Use of Render** — Deployed the complete React + FastAPI application as a single Web Service on Render.\n\n**Backboard** — Used Backboard as the family's memory layer to track recent meals and preferences.", "url": "https://wpnews.pro/news/what-s-for-dinner-aaj-kya-banega-a-20b-model-trained-to-solve-my-mom-s-kitchen", "canonical_source": "https://dev.to/iamani_0041/whats-for-dinneraaj-kya-banega-a-20b-model-trained-to-solve-my-moms-kitchen-crisis-without-the-3a8p", "published_at": "2026-10-04 19:09:46+00:00", "updated_at": "2026-10-04 19:12:11.236127+00:00", "lang": "en", "topics": ["large-language-models", "ai-tools", "ai-products", "generative-ai"], "entities": ["Tinker", "Backboard", "gpt-oss-20b", "Render", "FastAPI", "React", "mom-chef-v1", "Hacktoberfest"], "also_reported_by": [], "alternates": {"html": "https://wpnews.pro/news/what-s-for-dinner-aaj-kya-banega-a-20b-model-trained-to-solve-my-mom-s-kitchen", "markdown": "https://wpnews.pro/news/what-s-for-dinner-aaj-kya-banega-a-20b-model-trained-to-solve-my-mom-s-kitchen.md", "text": "https://wpnews.pro/news/what-s-for-dinner-aaj-kya-banega-a-20b-model-trained-to-solve-my-mom-s-kitchen.txt", "jsonld": "https://wpnews.pro/news/what-s-for-dinner-aaj-kya-banega-a-20b-model-trained-to-solve-my-mom-s-kitchen.jsonld"}}