📄 Resume: View Resume
- 🎓 I'm pursuing BTech in CSE with a major in Machine Learning
- 🔬 Currently exploring and building GPT-style transformer models
- ⚙️ Built a decoder-only transformer for spell correction using PyTorch
- 🔧 Customizing Karpathy’s minGPT and scaling via NanoGPT
- 🐞 Created a Bug Tracking System with role-based access, file uploads, and email alerts
- 💼 Interning at Hashbyte Studio, learning systems in gaming environments
- 🧭 Building my AI Agent Mastery Roadmap to reach top 1% LLM devs
- 🚀 Goal: Intern / freelance / contribute in ML within next 5 months
Transformer-based GPT decoder model trained to fix noisy sentences.
- Custom BPE tokenizer
- Implemented multi-head attention, feed-forward layers, and layer norm from scratch
- Trained in PyTorch on noisy-clean sentence pairs
- GitHub Repo | Kaggle Notebook
Role-based bug tracker built with Node.js, Express, MongoDB, React.
- Admin/Tester/Developer role flows
- Real-time email notifications
- File uploads and download routes
- Modular API structure
- Repo Link
Curated Notion/GitHub plan for mastering LLMs, AutoGPTs, and AI agents.
- Phase-based roadmap (Foundations → Projects → Agent Systems → Mastery)
- Includes lecture notes from CS224n, YouTube, blogs, and papers
- [In Progress]
- 💼 LinkedIn: linkedin.com/in/naman-rishy09
- 📊 Kaggle: kaggle.com/Rishy09
- 📨 Email: [namanchanana2005@gmail.com]
- Scaling SpellCorrectionGPT with NanoGPT
- Real-time AI agents and toolformer models
- Kaggle challenges (NLP, tabular ML)
- Self-hosted RAG pipelines
I don't just learn models, I break them, rebuild them, and make them work for real use-cases.