I don't just push boundaries; I gave them an eviction notice. Welcome to the lab.
Look, I don't just "code"—I teach rocks how to think and get them to thank me for it. I’m a founder-engineer who treats production like a jazz solo: calculated chaos with a lot of soul. I build high-ROI AI products because "impossible" is just a dare I haven't taken yet. My goal? Automate the mundane, scale the genius, and build the kind of tech that makes Skynet look like a pocket calculator. If you're looking for safe, color-by-numbers engineering, go hire a spreadsheet. I'm here to ship fast, break the status quo, and turn ungodly amounts of caffeine into pure, monetizable magic.
- 🔭 Currently Architecting: Axiom-One (Research-Grade Reasoning System—because chatbots are boring) & Agent-X (Vision-Language Browser Agent—yes, it can see you).
- 🌱 Downloading to Brain: Agentic AI, LangGraph, Fine-tuning... basically learning how to make the machines do my chores.
- 🤝 Looking for Co-Conspirators: To build innovative AI projects or anything that sounds vaguely impossible.
- 💬 Provoke Me About: System architecture, the singularity, or why AI ethics is the only conversation worth having at parties.
- ⚡ Weird Flex: I've fine-tuned 8 AI models, swept 3 hackathons, and type at 120 WPM. My blood stream is 90% Arabica.
|
Oct 2024 - Dec 2024 AI agent that automates browser tasks using vision-language models. Interacts with websites from plain English instructions, leveraging Python, FastAPI, Playwright, React, and Gemini 2.5. Open-source, safety-first, and fully browser-native for hands-free web automation.
|
Jan 2025 - Present A visual debugger for AI agent loops. Add
|
|
Jun 2024 - Sep 2024 Comprehensive RAG course covering theory, production-ready code, and 15 modules—reranking, chunking, embeddings, vector DBs, hybrid search, query optimization, agentic workflows, and deployment. Suited for AI/ML developers seeking advanced RAG knowledge.
|
Aug 2024 - Dec 2024 A Research-Grade Reasoning System (Not a Chatbot). A thinking-first research agent that decomposes problems, runs multi-step reasoning loops, validates its own outputs, and exposes the entire thought pipeline. Optimizes for correctness under pressure.
|
- 🎓 Building toward Computer Use with Anthropic - Anthropic (Mar 2025)
- 🏢 Software Engineering Job Simulations - EA, Walmart, Goldman Sachs (Forage, 2024)
- 🤖 AI Agents in LangGraph - DeepLearning.AI (2024)
- 🚀 Finetuning Large Language Models - DeepLearning.AI (2024)
- 🎨 Front End Development Libraries - freeCodeCamp (Feb 2025)
- 👁️ Prompt Engineering for Vision Models - DeepLearning.AI (Dec 2024)
- 🧠 Machine Learning with Python - Cognitive Class (Dec 2024)
- 🔗 Blockchain Masterclass - CFTE
- 💼 LinkedIn: Ravani Roshan (My "Professional" Mask)
- 🌐 Portfolio: ravani-roshan-singh.vercel.app/certificates (The "I'm Actually Good" Proof)
- 📧 Email: ravaniroshansingh@gmail.com (Send memes or job offers. Preferably both.)
- 💻 HackerRank: Profile
"I don't just solve problems. I create solutions that make the original problem look like it was asking for it." 💪
Fair warning: I may accidentally automate your job while we're talking. 🌟