Transforming Business Through AI
I am a PhD computer scientist turned AI consultant, and founder of Chelsea AI Ventures. I've authored several books including Amazon bestsellers Generative AI with LangChain and Machine Learning for Time-Series with Python, the Artificial Intelligence with Python Cookbook, and -- most recently -- a book about RAG that covers both fundamentals and state-of-the-art.
I transform businesses through enterprise-scale AI solutions that deliver measurable impact. From fraud detection systems saving millions to real-time decision engines processing 100k+ daily transactions, I bring deep technical expertise and practical business focus to every project.
- Built enterprise-scale fraud detection systems preventing millions in annual losses
- Developed real-time ML decision engine processing 100k+ daily transactions at <300ms latency
- Improved marketing ROI by 15% through ML-driven optimization at a major travel platform
- Published bestselling books on GenAI and ML implementation with 100k+ in sales
The research from the time of my PhD (KTH, CS) has been cited close to 600 times, and for my innovation is combining quick-wins and strategic, more experimental, sometimes cutting-edge work that carries longer-term payoffs. I combine technical expertise with practical strategy to help organizations succeed in their AI journey.
AI Consulting
Education
Next public workshop
The GenAI Build Lab: Build Production-Ready RAG with Open Models
Hosted by Packt Publishing
Most RAG tutorials stop at the first plausible answer. This lab starts there. On a travel-company case study you build a RAG system on small open models, expose its failure modes on purpose, fix retrieval with hybrid search and reranking, add corrective retrieval and answer self-checking, benchmark quality against latency and cost, then put open guardrail models in front of the finished app. It runs without a paid API key, a GPU cluster, or a per-token bill.
- Hybrid keyword-plus-dense retrieval, Reciprocal Rank Fusion, cross-encoder reranking, MMR
- Corrective retrieval and answer self-checking
- RAGAS and retrieval metrics, plus a latency, quality and cost scorecard
- Llama Guard and Granite Guardian as guardrails
- Runnable notebooks, reranking and guardrail templates, and a production-readiness checklist you keep
$119.99 with the speaker discount, 40% off the $199.99 ticket
Book a ticketPackt Publishing has hosted 217 events for 10.6k attendees. Includes a certificate of completion and the full recording.
If you're a founder or technical leader weighing build vs. buy on an AI initiative, message me to set up a 30-minute scoping call.