Ram Murmu
Full Stack AI Developer β’ Product Designer β’ RunAsh AI
Visit RunAsh
- About Me
- Highlights
- Modern Tech Stack
- Featured Projects
- AI Models & Datasets
- Experience & Features
- How to Collaborate
- Let's Connect
- License
Name: Ram Murmu
Title: AI Innovator | Full Stack AI Developer | Product Designer
Company: RunAsh AI
Location: π Global (MIT, USA)
Interests:
- LLMs
- AI UI
- Live Streaming
- Human-Centric AI
- Open Source
Contact: rammurmu@outlook.in
Website: https://rammurmu.comI build at the intersection of AI and modern user experience β shipping production-ready models and delightful interfaces for live, interactive systems.
- Founder & CEO at RunAsh AI β live streaming + AI for interactive experiences.
- Author of RunAsh Language Model (RALM) and multiple Hugging Face model cards.
- Experience building real-time CV and NLP systems (PyTorch, Transformers, OpenCV).
- Open-source maintainer and Kaggle competitor.
- Languages: Python, TypeScript, JavaScript
- Frameworks: React, Next.js, FastAPI, Node.js
- AI/ML: PyTorch, TensorFlow, Transformers, Scikit-learn, OpenCV
- Cloud / Infra: AWS, Vercel, Docker
- Design: Figma, TailwindCSS, Framer Motion
| Project | Description | Tech |
|---|---|---|
| RunAsh AI | AI-powered live streaming platform (real-time chat + CV integrations) | Next.js, React, PyTorch, FastAPI |
| RunAsh Chat | Conversational AI for live streaming (moderation, summarization, emotes) | Transformers, WebSocket, React |
| Sentiment Behaviour Support Bot | Mental health & sentiment-aware support chatbot | Python, FastAPI, Hugging Face |
| Live Streaming Chat Dataset (LSCD) | 10,000+ annotated samples curated for stream chat behaviour | Dataset (Hugging Face) |
- RunAsh Language Model (RALM): fine-tuned LLM for live-stream chat agents β https://huggingface.co/RamMurmu/RunAsh-Language-Model
- Live Streaming Chat Dataset (LSCD): dataset used to fine-tune chat behaviour models β https://huggingface.co/datasets/RamMurmu/Live-Streaming-Chat-Dataset
- Real-time systems: WebSocket-based low-latency chat, inference pipelines, and online model serving.
- CV for streams: face & object detection for overlays and interaction triggers (OpenCV, YOLO-style detectors).
- UX-first AI: human-centered prompts, fallback strategies, and safe-moderation layers.
- Production ML: training pipelines, CI/CD, Dockerized inference, metrics & monitoring.
Looking to collaborate? Here are a few ways:
- Open-source: raise an issue or PR on any repo β I'm happy to review.
- Research & partnerships: email me at rammurmu@outlook.in or reach via LinkedIn.
- Product/Startup: interested in integrating AI into streaming or interactive experiences? Let's discuss: https://runash.in
- Portfolio: https://rammurmu.com
- Hugging Face: https://huggingface.co/rammurmu
- LinkedIn: https://linkedin.com/in/rammurmu
- X (Twitter): https://x.com/rammurmuu
- Kaggle: https://www.kaggle.com/rammurmu
- Email: rammurmu@outlook.in
More About Me
- Currently at MIT, accelerated by Y Combinator, building for the future of live interactive AI.
- Kaggle competitor, Hugging Face model author, open source enthusiast.
- If you'd like a custom README version (different banner, colors, or focus areas), tell me what tone and sections to prioritize and I'll create a tailored variant.