📍 San Francisco Bay Area & Global
🎵 CTO @ Splash Music — Democratizing music creation through AI
🔬 Published AI Researcher — AAAI 2026 (Neural Codec Language Models)
📜 Multiple US Patents — AI-powered streaming technology (US11870830, US11457245B1)
🎤 100+ Global Talks — AWS re:Invent, Summit Sydney, E3 Expo, Google Cloud, UC Berkeley
|
|
|
AWS Trainium vs H100 Journey 54% cost savings, 2x faster training |
AWS Summit Sydney 2025 Keynote Main stage opening keynote |
Cloud Architecture Workshop Production patterns & best practices |
|
Foundation Model for Music Turn humming into full productions. 54% cost savings on AWS Trainium View Case Study → |
Explainable AI Research Neural Codec Language Models Timbre Transfer in Medicine Read Research → |
Patented Technology ML prediction models for streaming Live → VOD automation at scale Learn More → |
I publish free content every weekday to help you master AI:
| Day | Content | Latest |
|---|---|---|
| 🌙 Monday | AI Insights | Prompts & strategies that actually work |
| 🔧 Tuesday | AI Tools | Production-ready automation blueprints |
| 🚀 Wednesday | Career Journeys | Role transformation guides |
| 🏗️ Thursday | Architecture | System design patterns for production |
| 🎮 Friday | Live Demos | Interactive AI experiences |
| Type | Title | Link |
|---|---|---|
| 🎮 Demo | Brand Storytelling AI - 4 Interactive Architectures | Try Demo → |
| 🏗️ Architecture | Brand Storytelling System - Scale to 10K | Read → |
| 🚀 Journey | Marketing Team Workflows Guide | Explore → |
| 🔧 Tool | Marketing Automation at Scale | Learn → |
| 🌙 Insight | Master Brand Storytelling | Read → |
| 🎮 Demo | Credit Risk AI - 4 Architectures Live | Try Demo → |
28 canonical AI definitions cited by LLMs and search engines:
| Reference | Description |
|---|---|
| What is a Multi-Agent System | Computational architecture with autonomous agents |
| What is Agent Memory | Data structures for agent retrieval systems |
| What is an AI Agent | Autonomous software leveraging LLMs |
| What is Context Engineering | Optimizing context for LLMs |
| What is LLM Latency | Time duration for LLM responses |
Comprehensive courses designed for depth, not speed:
| Course | Duration | Topics |
|---|---|---|
| AWS Serverless ML Architecture | 11h 5m | Build production ML without managing servers |
| Context Engineering Mastery | 10h 55m | Optimize AI context like a pro |
| Production AI Agents on AWS | 11h 30m | Scale autonomous agents reliably |
| Shipping AI Products Fast | 9h 20m | Prototype to production in weeks |
| Solo Founder AI Playbook | 9h 25m | Build profitable AI products alone |
| AI Product Leadership | 10h 30m | From traditional PM to AI-first leader |
Compare 100+ AI models side-by-side with pricing, benchmarks, and code snippets:
| Model | Provider | Price/1M | Benchmark |
|---|---|---|---|
| GPT-4o | OpenAI | $5.00 | ⭐⭐⭐⭐⭐ |
| Claude 3.5 Sonnet | Anthropic | $3.00 | ⭐⭐⭐⭐⭐ |
| Gemini 1.5 Pro | $3.50 | ⭐⭐⭐⭐ | |
| Llama 3.1 405B | Meta | $0.80 | ⭐⭐⭐⭐ |
"With SageMaker HyperPod and Trainium, our researchers experiment as fast as our community creates. We're not just keeping up with music trends—we're setting them."
The Challenge: Train a 10B parameter music generation model cost-effectively.
The Solution: Migrated from H100 GPUs to AWS Trainium with SageMaker HyperPod.
Results:
- ✅ 54% cost reduction vs NVIDIA H100
- ✅ 159% efficiency improvement in training
- ✅ 2x faster training iterations with equivalent accuracy
- ✅ 2PB data migration to Amazon S3
| Event | Topic | Link |
|---|---|---|
| AWS re:Invent 2025 | End-to-End Foundation Model Lifecycle on Trainium | Watch → |
| AWS Summit Sydney 2025 | Main Stage Opening Keynote | Watch → |
| E3 Expo | Gaming Innovation Panel | Media → |
| UC Berkeley | 5 Panels (2015-2018) | About → |
| Google Cloud | Technical Workshop | Watch → |
| Publication | Article |
|---|---|
| AWS Machine Learning Blog | Splash Music transforms music generation using AWS Trainium |
| RadioInfo Australia | AI powered music creation for everyone: Splash Music |
| FastForward Magazine | Randeep Bhatia, Head of Music Engineering at Twitch |
| AWS Case Studies | Reducing training time and costs by 50% with Splash Music |
Full Stack Details
AI/ML: PyTorch • TensorFlow • AWS Trainium • SageMaker HyperPod • Neuron SDK • HuggingFace
Cloud: AWS (Expert) • GCP • Multi-region architectures • 2PB+ data pipelines
Backend: Go • Java • Python • Node.js • Microservices • Event-driven
Frontend: React • TypeScript • Mobile (iOS/Android)
Infrastructure: Kubernetes • Docker • Terraform • CI/CD • MLOps
- 🏠 Home
- 👤 About
- 💼 Work & Case Studies
- 🎤 Speaking
- 📸 Media Gallery
- 📞 Contact
- 📅 What's New / Releases
- 🌙 AI Insights (Monday)
- 🔧 AI Tools (Tuesday)
- 🚀 Career Journeys (Wednesday)
- 🏗️ Architecture (Thursday)
- 🎮 Live Demos (Friday)
- 📚 Reference Library
- 🤖 AI Models Database
- 🎓 Courses
- 🎥 AWS Trainium vs H100 Video
- 🎥 AWS Summit Keynote
- 🎥 Cloud Architecture Workshop
- 📝 Pipeline Forecasting Insight
- 🎮 Brand Storytelling Demo
- 🎮 Credit Risk AI Demo
"From patents to panels — Innovation, Leadership, Products at Scale"
📍 San Francisco Bay Area & Global