Recommender Systems
We build smart recommendation engines that help people find what they love — whether it’s movies, products, or content.
- • MovieLens/TMDB demos, e-commerce ready
- • Cold-start & multi-task CTR/CVR
- • Batch + real-time serving
AI built with clarity.
Calm, clear, outcome-driven machine learning. Breathe. Build. Ship.
Mindful, production-ready AI. We ship fast, stable, and simple—without drama.
We build smart recommendation engines that help people find what they love — whether it’s movies, products, or content.
Task-focused agents that ingest, clean, learn, and act—so you don’t have to.
Train, track, deploy on GCP/AWS/Azure. Minimal ops, maximal reliability.
One-week sprints to validate ideas. From notebook to clickable demo.
Architecture reviews, roadmaps, and hands-on pairing to upskill your team.
Your data, your edge—vertical models, private fine-tunes, and on-prem options.
Hands-on demos you can click and try. Built fast, stable, and simple—so you can feel the value, not just read about it.
Rate a few films and get instant recommendations. ALS/ItemKNN plus DCNv2/MMoE variants.
Point at a CSV or DB; it profiles, cleans, and trains a quick model with a readable report.
Script-to-shots prototype: generates boards, voices, and rough cuts for a concept reel.
An ops assistant that deploys, monitors, and reports on your ML jobs and costs.
Vertex AI/Kubeflow pipeline template with experiment tracking and model registry.
Interactive dashboard for NDCG@K, CTR/CVR, coverage, diversity, and more.

We build mindful, production-ready AI — prototypes, recommender systems, and intelligent agents. My goal is to help startups validate ideas fast without burning time or money.
With a background in cloud ML engineering (Google Cloud Professional ML Engineer), we combine technical depth with mindfulness — clarity and simplicity in every project.
Anapana.ai helps startups validate ideas quickly with prototypes, recommender systems, and custom AI agents — fast, simple, and production-ready.