

Practical machine learning, forecasting and AI evaluation
I help product teams decide whether a model is useful for their actual workflow. That might mean an evaluation set for a support assistant or a forecast with honest error bounds. You receive reproducible notebooks, documented assumptions and a results walkthrough, including the cases where the approach does not work well.



We had a tight launch window. The work on error analysis was ready for our team to review on schedule.
Service: I will evaluate your AI assistant against real support questions
This was our first time outsourcing this work. The evaluation dataset and the patient explanations made it easy.
Service: I will evaluate your AI assistant against real support questions
Seller's reply
Thank you for the review, and for sharing your numbers afterwards — that's genuinely useful for me.
Our brief was understood from the first draft. The evaluation dataset made the whole project worthwhile.
Service: I will evaluate your AI assistant against real support questions
This was our first time outsourcing this work. The error analysis and the patient explanations made it easy.
Service: I will evaluate your AI assistant against real support questions
Seller's reply
Thanks, it was a pleasure. Still around if anything comes up.
The error analysis matched our references closely, and the final files were organised well.
Service: I will evaluate your AI assistant against real support questions
Really pleased with the evaluation dataset. We could put the delivery to use without another meeting.
Service: I will evaluate your AI assistant against real support questions
Seller's reply
Thank you — your brief was clear, which made the whole thing easier. Message me any time.